Kinds of data

Audio and speech

Recorded speech: conversations, calls, interviews, podcasts and lectures, in any language.

Public price: $65 to $266
per hour of audio
Demand: Rising
Who buys: Voice and speech model companies, labs, speech data vendors

The market rate for audio, by side

The median where three or more figures exist, otherwise the figures we hold. Each side is a different point in the market: what the people who make it are paid, what holders are offered, what catalogues ask, what was paid.

Paid to the people who make it

$24

per hour of work, median of 31 figures; middle half $14 to $42

Paid to the people who make it

$30

per hour of audio, median of 3 figures; middle half $21 to $115

Sales that happened

$76

per hour of audio, median of 9 figures; middle half $40 to $2,005

Asking prices off the shelf

$118

per hour of audio, median of 109 figures; middle half $65 to $266

Asking prices off the shelf

$1.27 to $1.49

per record, from 2 figures

Made to order

$122

per hour of audio, median of 9 figures; middle half $90 to $245

Cost to produce

$244 to $333

per hour of audio, from 2 figures

What audio has sold or been offered for

The whole price board

Public prices for audio
What was sold or offered	Price	When	What to know	Source
Custom speech data collection to a buyer's spec (recruited speakers, recorded, transcribed and QA'd), 4 to 12 weeks	$50 to $300 per hour of delivered audio per hour of delivered audio	Oct 2026	AIxBlock's range after transcription and QA, set by language scarcity, accent diversity and annotation depth; the same range appears in its call-center guide (held under aixblock-custom-1000-hours). The page also warns that quotes below $20 an hour for custom collection usually signal unsustainable contributor rates. The page carries no date (footer 2025).	AIxBlock, How to Choose a Speech Data Collection Service for ASR A vendor's published price guide
Studio-quality text-to-speech (TTS) voice data recorded to order	$200 to $500 per hour per hour of audio	Oct 2026	AIxBlock puts studio-quality TTS data above its $50 to $300 range for ordinary custom speech collection. The page carries no date (footer 2025).	AIxBlock, How to Choose a Speech Data Collection Service for ASR A vendor's published price guide
Custom on-site speech and audio recordings made to the buyer's requirements by field teams (Spire Light on Datarade)	Usage-based $250 / Audio hour per hour of audio recorded to order	Oct 2026	A 'starts at' price on a data marketplace; Spire Light is the company behind SpeechData.ai (its footer reads 'A Spirelight company'), and detailed pricing is by request. Delivered as .json and .wav with human QA.	Datarade, Custom On-Site Speech & Audio Data Collection (Spire Light) A catalogue or marketplace listing
Custom paired conversational speech recordings (audio, optionally video) built to project requirements in 80+ languages (Spire Light on Datarade)	Usage-based $90 / Audio hour per hour of audio recorded to order	Oct 2026	A 'starts at' price on a data marketplace for paired conversations recorded to order from a network of over 30,000 people; the $250 on-site listing from the same seller is for field recordings. Detailed pricing is by request.	Datarade, Custom Pair Conversational Speech Audio / Video Data Collection Dataset in 80+ Languages (Spire Light) A catalogue or marketplace listing
Median advertised hourly pay for voice acting and voice-over roles in AI training	$35 an hour (live median) per hour, the median of 111 advertised rates	Oct 2026	The median of 111 distinct advertised hourly rates across 8 platforms; the typical range is $35 to $75 and the top is $150 at Mercor. AITraining.jobs has the same owner as AIData.Broker.	AITraining.jobs, Voice Acting and Voice Over Jobs for AI A catalogue or marketplace listing
Nexdata Indonesian spontaneous dialogue smartphone speech, 747 hours from 1,190 native speakers, transcribed with PII redacted (on Datarade)	Starts at $20,000 / purchase (one-off) per dataset of 747 hours	Oct 2026	A 'starts at' one-off licence price, the same $20,000 floor Nexdata shows across its Datarade speech products; spontaneous two-party dialogue at this floor works out cheaper per hour than Nexdata's scripted Indonesian set.	Datarade, Nexdata Indonesian Spontaneous Dialogue Smartphone Speech Dataset, 747 Hours A catalogue or marketplace listing
Nexdata Indonesian scripted monologue smartphone speech, 359 hours from 496 speakers, transcribed (on Datarade)	Starts at $20,000 / purchase (one-off) per dataset of 359 hours	Oct 2026	A 'starts at' one-off licence price; Nexdata shows the same $20,000 floor on many of its Datarade speech products, so it reads as a minimum order rather than a priced quote. The same corpus lists on ELRA at EUR 57,978.50 for commercial use, about 2.6 times this floor.	Datarade, Nexdata Indonesian Speech Data by Mobile Phone, 359 Hours A catalogue or marketplace listing
REMI Data licensed native-speaker speech recordings in 13+ languages including Yoruba, Amharic and Swahili, WAV 16 kHz with verbatim transcripts and perpetual commercial consent (on Datarade)	Usage-based $120 / per hour (one-off purchase from $500) per hour of audio	Oct 2026	One per-hour rate across all the languages the listing names, Yoruba among them; the stated volume is 100 hours for the whole product, so how many Yoruba hours exist off the shelf is not stated, and custom collection is offered beside it.	Datarade, REMI Data Multilingual Speech Recordings, 13+ Languages A catalogue or marketplace listing
IARPA Babel Vietnamese Language Pack IARPA-babel107b-v0.7 (LDC2017S01): about 201 hours of Vietnamese conversational and scripted telephone speech with transcripts, collected by Appen in 2012	$25.00 (non-member fee) per corpus of about 201 hours	Oct 2026	LDC's non-member fee, shown under View Fees (the reduced-licence fee is also $25.00). The Babel packs were paid for by the US government and are priced near zero; non-member LDC licences exclude commercial use, so this is a research price, not a market price for Vietnamese speech.	LDC catalog, IARPA Babel Vietnamese Language Pack IARPA-babel107b-v0.7 (LDC2017S01) A catalogue or marketplace listing
NaijaVoices suite of speech-text datasets (Hausa, Igbo and Yoruba; 1,800+ hours from 5,000+ speakers): the Supporter tier that adds a waiver for commercial use	Donate $5,000 or More per commercial-use waiver for the suite of 1,800+ hours	Oct 2026	A community project's minimum donation for commercial rights, not a negotiated price; the default licence (CC BY-NC-SA 4.0) is free for non-commercial use. Read speech of expert-written sentences, so it sits far below conversational listings.	NaijaVoices, Membership tiers A catalogue or marketplace listing
NaijaVoices: total cost of creating its 1,800+ hour speech-text dataset in Hausa, Igbo and Yoruba	substantial (>$600k) per dataset of 1,800+ hours	Oct 2026	The producer's own statement of what the dataset cost to make (sentence writing, recording, double review), set against its $5,000 commercial waiver: the commercial right sells at about 1 percent of production cost.	NaijaVoices, Membership FAQ The payer's own page
NaijaVoices: the rate it costs per hour of recording by its Nigerian 'data farmers' (Hausa, Igbo, Yoruba read speech)	rate of $30/hour for recording per hour of recording	Oct 2026	The rate NaijaVoices uses to turn a donation into hours of speech data; the page also states $0.12 per sentence written, a double review of all output, each sentence recorded twice and at most 5 hours per speaker.	NaijaVoices, Membership: donation impact calculation breakdown The payer's own page
SpeechData.ai Indonesian conversational speech dataset: 2,000 hours of spontaneous two-person conversations recorded in Indonesia, 200 native speakers, dual-channel, time-aligned transcripts	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf asking price for a commercial training and evaluation licence with speaker consent documentation; the page says volume discounts are available for full-dataset and multi-language purchases.	SpeechData.ai, Indonesian conversational speech dataset A catalogue or marketplace listing
SpeechData.ai Vietnamese conversational speech dataset: 2,000 hours of spontaneous two-person conversations recorded in Vietnam, 200 native speakers, dual-channel, time-aligned transcripts	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf asking price for a commercial training and evaluation licence with speaker consent documentation; the page says volume discounts are available for full-dataset and multi-language purchases.	SpeechData.ai, Vietnamese conversational speech dataset A catalogue or marketplace listing
Off-the-shelf speech data, major languages	$50-100 per hour of audio per hour	Oct 2026	A vendor's price table; low-resource languages are $60 to $150 on the same table, custom collection $150 to $400+.	SpeechData.ai, How to buy speech data (price table) A vendor's published price guide
Recorded speech sold off the shelf	$60-150 per hour of audio per hour	Oct 2026	A vendor's account of the market for off-the-shelf conversational speech; it raised the floor from $25 (read 11 October 2026) to $60. Its own catalogue sits at $60 to $95.	SpeechData.ai, How to buy speech data (FAQ) A vendor's published price guide
AI Tutor specialised in multilingual audio, Yoruba (xAI): records, annotates and transcribes Yoruba speech and localises English text into Yoruba	$35/hour - $45/hour per hour (paid to the tutor, US-based candidates)	Oct 2026	The rate is stated for US-based candidates only; the post says international candidates get their rate during recruitment, so a Yoruba speaker in Nigeria is likely paid differently (not stated). Same US band as xAI's other AI Tutor posts (French, Hausa). Posted 2026-08-07, updated 2026-10-13; contractor hours vary, and projects may require at least 10 hours a week.	xAI job board (Greenhouse), AI Tutor - Yoruba The payer's own page
Custom 1,000-hour English call-center speech collection with accent coverage and basic transcription, 12 weeks	$80,000 to $200,000 per 1,000-hour project	Oct 2026	Same guide prices custom collection at $50 to $300 per delivered hour, versus $5,000 to $25,000 for the same volume off the shelf.	AIxBlock A vendor's published price guide
Off-the-shelf call-center audio licensed from existing corpora (AIxBlock guide)	$5 to $25 per hour of audio per hour of audio	Oct 2026	Undated vendor guide read on this date; English call-center audio sits at the low end, low-resource languages and regulated domains at the high end.	AIxBlock A vendor's published price guide
Sound-effects catalogue for AI training: 350+ libraries, 53,000 files (Krampfstadt Studio), one-off purchase	€60,000 / purchase per catalogue of 53k sound files	Oct 2026	The same seller asks $78,900 for this catalogue on Opendatabay; a yearly licence is EUR 20,000 and a monthly one EUR 3,000.	Datarade, Krampfstadt Studio 350+ SFX libraries, 53k audio files A catalogue or marketplace listing
Sound-effects catalogue for AI training: 350+ libraries, 53,000 files (Krampfstadt Studio), yearly licence	€20,000 / year per year	Oct 2026	Three years of the yearly licence cost about what the one-off purchase does (EUR 60,000).	Datarade, Krampfstadt Studio 350+ SFX libraries, 53k audio files A catalogue or marketplace listing
Two-speaker spontaneous conversations collected on demand in rare and low-resource languages and accents, with transcripts (Silencio Network)	Usage-based $50 / Hour per hour of audio collected to order	Oct 2026	Collection to order from a crowd of contributors in 180+ countries, priced per delivered hour, with discounts above 1,000 hours.	Datarade, Silencio Network Spontaneous Conversations On-Demand (Accent and Dialect Focus) A catalogue or marketplace listing
Real customer-service calls in English and Hindi for chatbot and voicebot training (WiserBrand), usage-based licence	Usage-based $25 / hour per hour of audio (usage-based licence)	Oct 2026	Same per-hour price as the seller's US English calls: no premium for Hindi or code-switched speech at this seller.	Datarade, WiserBrand Chatbot Conversation Dataset (English and Hindi) A catalogue or marketplace listing
Real US customer-service call recordings, speaker-separated 48 kHz, PII-redacted (WiserBrand, 40,000+ hours available), usage-based licence	Usage-based $25 / hour (one-off purchase from $10,000) per hour of audio (usage-based licence)	Oct 2026	Seller's own per-hour list price on the Datarade marketplace for real (not staged) support calls; a one-off purchase starts at $10,000.	Datarade, WiserBrand Real Call Center Conversations for ASR and Voice AI A catalogue or marketplace listing
ALLIES corpus: 900 hours of French broadcast news for diarization evaluation (ELRA catalogue ELRA-S0486)	31250.00 € (non-member, commercial use) per corpus of 900 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Broadcast speech in volume, so cheap per hour. A non-commercial-use licence is EUR 2,500 for an academic and EUR 9,375 for a company.	ELRA catalogue, ALLIES Corpus (ELRA-S0486) A catalogue or marketplace listing
American children's read speech (storybooks, songs), microphone, 50.5 hours from 219 children (ELRA catalogue ELRA-S0468)	28785.00 € (non-member, commercial use) per corpus of 50.5 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech: about three and a half times the per-hour price of the adult American English set in the same series (ELRA-S0467). ELRA lists no non-commercial licence for it.	ELRA catalogue, American Children Speech Data by Microphone - 50 Hours (ELRA-S0468) A catalogue or marketplace listing
American English prompted speech recorded on mobile phones, 800 hours from 1,842 speakers (ELRA catalogue ELRA-S0437)	136800.00 € (non-member, commercial use) per corpus of 800 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.	ELRA catalogue, American English Speech Data by Mobile Phone - 800 Hours (ELRA-S0437) A catalogue or marketplace listing
American English read speech recorded on mobile phones, 215 hours from 349 speakers (ELRA catalogue ELRA-S0467)	34722.50 € (non-member, commercial use) per corpus of 215 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, American English Speech Data by Mobile Phone_Reading - 215 Hours (ELRA-S0467) A catalogue or marketplace listing
Italian Speech Corpus 1 (Appen): 44 hours from 202 speakers, 4 microphone channels, office and public places (ELRA catalogue ELRA-S0147)	15000.00 € (non-member, commercial use) per corpus of 44 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. An Appen off-the-shelf corpus distributed through ELRA, one of the few Appen corpora with a public price. A non-commercial-use licence is EUR 1,500 for an academic and EUR 15,000 for a company.	ELRA catalogue, Italian Speech Corpus 1 (Appen) (ELRA-S0147) A catalogue or marketplace listing
Spanish Speech Corpus 1 (Appen): 39 hours from 200 speakers, 4 microphone channels, office and public places (ELRA catalogue ELRA-S0149)	15000.00 € (non-member, commercial use) per corpus of 39 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. An Appen off-the-shelf corpus distributed through ELRA. A non-commercial-use licence is EUR 1,500 for an academic and EUR 15,000 for a company.	ELRA catalogue, Spanish Speech Corpus 1 (Appen) (ELRA-S0149) A catalogue or marketplace listing
Brazilian Portuguese prompted speech recorded on mobile phones, 1,044 hours from 2,038 speakers (ELRA catalogue ELRA-S0445)	247950.00 € (non-member, commercial use) per corpus of 1,044 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.	ELRA catalogue, Brazilian Portuguese Speech Data by Mobile Phone - 1,044 Hours (ELRA-S0445) A catalogue or marketplace listing
BREF-120: 100 hours of French read speech (ELRA catalogue ELRA-S0067)	15000.00 € (non-member, commercial use) per corpus of 100 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. A classic read-speech corpus. A non-commercial-use licence is EUR 4,000 for an academic and EUR 15,000 for a company.	ELRA catalogue, BREF-120 - A large corpus of French read speech (ELRA-S0067) A catalogue or marketplace listing
British children's read speech (textbooks, storybooks), high-fidelity microphone, 55 hours from 201 children (ELRA catalogue ELRA-S0474)	39187.50 € (non-member, commercial use) per corpus of 55 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech: about three times the per-hour price of the adult British English set in the same series (ELRA-S0466). ELRA lists no non-commercial licence for it.	ELRA catalogue, British Children Speech Data by Microphone - 55 Hours (ELRA-S0474) A catalogue or marketplace listing
British English prompted speech recorded on mobile phones, 831 hours from 1,651 speakers (ELRA catalogue ELRA-S0448)	213151.50 € (non-member, commercial use) per corpus of 831 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.	ELRA catalogue, British English Speech Data by Mobile Phone - 831 Hours (ELRA-S0448) A catalogue or marketplace listing
British English read speech recorded on mobile phones, 199 hours from 346 speakers (ELRA catalogue ELRA-S0466)	47262.50 € (non-member, commercial use) per corpus of 199 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, British English Speech Data by Mobile Phone_Reading - 199 Hours (ELRA-S0466) A catalogue or marketplace listing
Cantonese Speecon: about 231 hours of adult and child speech over 4 microphone channels in office, car and public places (ELRA catalogue ELRA-S0287)	75000.00 € (non-member, commercial use) per corpus of 231 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Multi-microphone, multi-environment recordings. A non-commercial-use licence is EUR 60,000 for an academic and EUR 75,000 for a company.	ELRA catalogue, Cantonese Speecon database (ELRA-S0287) A catalogue or marketplace listing
Chinese Kids Speech database (Lower Grade): 87 hours from 184 children aged 6 to 10, smartphone (ELRA catalogue ELRA-S0496)	15400.00 € (non-member, commercial use) per corpus of 87 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech. A non-commercial-use licence is EUR 2,600 for an academic and EUR 10,300 for a company.	ELRA catalogue, Chinese Kids Speech database (Lower Grade) (ELRA-S0496) A catalogue or marketplace listing
Chinese Kids Speech database (Upper Grade): 72 hours from 161 children aged 10 to 12, smartphone (ELRA catalogue ELRA-S0497)	14000.00 € (non-member, commercial use) per corpus of 72 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech. A non-commercial-use licence is EUR 2,300 for an academic and EUR 9,300 for a company.	ELRA catalogue, Chinese Kids Speech database (Upper Grade) (ELRA-S0497) A catalogue or marketplace listing
CLE Pakistan Urdu Speech Corpus: 61.62 hours of phonetically rich read Urdu on varied microphones (ELRA catalogue ELRA-S0403)	23400.00 € (non-member, commercial use) per corpus of 61.62 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Read Urdu here costs several times conversational Urdu in commercial catalogues. A non-commercial-use licence is EUR 15,600 for an academic and EUR 23,400 for a company.	ELRA catalogue, CLE Pakistan Urdu Speech Corpus (ELRA-S0403) A catalogue or marketplace listing
Corpus of Spontaneous Japanese (CSJ), 650 hours as listed by ELRA, distributed for NINJAL (ELRA catalogue ELRA-S0488)	43125.00 € (non-member, commercial use) per corpus of 650 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. The owner, NINJAL, licenses the same corpus directly for 5 million yen (open-ended, up to 10 users), well under ELRA's price. A non-commercial-use licence is EUR 431.25 for an academic and EUR 43,125 for a company.	ELRA catalogue, Corpus of Spontaneous Japanese (CSJ) (ELRA-S0488) A catalogue or marketplace listing
Danish SpeechDat(II) FDB-4000: 242 hours of fixed-network telephone speech from 4,000 speakers (ELRA catalogue ELRA-S0073)	56000.00 € (non-member, commercial use) per corpus of 242 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Fixed-network telephone speech from 4,000 speakers. A non-commercial-use licence is EUR 48,000 for an academic and EUR 56,000 for a company.	ELRA catalogue, Danish SpeechDat(II) FDB-4000 (ELRA-S0073) A catalogue or marketplace listing
Dutch Polyphone Database: 335 hours of Dutch and Flemish telephone speech (ELRA catalogue ELRA-S0010)	35000.00 € (non-member, commercial use) per corpus of 335 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Telephone speech. A non-commercial-use licence is EUR 20,000 for an academic and EUR 35,000 for a company.	ELRA catalogue, Dutch Polyphone Database (ELRA-S0010) A catalogue or marketplace listing
Egyptian Arabic Speecon: 196 hours of Modern Standard Arabic as spoken in Egypt, adults and children, 4 microphone channels in office, car and public places (ELRA catalogue ELRA-S0308)	75000.00 € (non-member, commercial use) per corpus of 196 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Multi-microphone, multi-environment recordings. A non-commercial-use licence is EUR 60,000 for an academic and EUR 75,000 for a company.	ELRA catalogue, Egyptian Arabic Speecon database (ELRA-S0308) A catalogue or marketplace listing
ESTER 2: about 100 hours of transcribed French broadcast news (ELRA catalogue ELRA-S0338)	25000.00 € (non-member, commercial use) per corpus of 100 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Commercial use costs several times the non-commercial price. A non-commercial-use licence is EUR 2,000 for an academic and EUR 7,500 for a company.	ELRA catalogue, ESTER 2 Corpus (ELRA-S0338) A catalogue or marketplace listing
EthioSpeech: 391 hours of read speech in six Ethiopian languages (Amharic, Tigrigna, Oromo, Somali, Afar, Sidama) (ELRA catalogue ELRA-S0494)	5400.00 € (non-member, commercial use) per corpus of 391 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Low-resource African languages at a low commercial price per hour. A non-commercial-use licence is free for an academic and EUR 5,400 for a company.	ELRA catalogue, EthioSpeech (ELRA-S0494) A catalogue or marketplace listing
French prompted speech recorded on mobile phones, 769 hours from 1,623 speakers (ELRA catalogue ELRA-S0446)	182637.50 € (non-member, commercial use) per corpus of 769 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.	ELRA catalogue, French Speech Data by Mobile Phone - 769 Hours (ELRA-S0446) A catalogue or marketplace listing
French read speech recorded on mobile phones (France, Canada, Africa), 231.9 hours from 406 speakers (ELRA catalogue ELRA-S0465)	50670.15 € (non-member, commercial use) per corpus of 231.9 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, French Speech Data by Mobile Phone_Reading - 231 Hours (ELRA-S0465) A catalogue or marketplace listing
English spoken by native German speakers (accented English), mobile phones, 535 hours from 1,162 speakers (ELRA catalogue ELRA-S0442)	152475.00 € (non-member, commercial use) per corpus of 535 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Accented English priced at the top tier, the same as native German speech in the series. ELRA lists no non-commercial licence for it.	ELRA catalogue, German Speaking English Speech Data by Mobile Phone - 535 Hours (ELRA-S0442) A catalogue or marketplace listing
German prompted speech recorded on mobile phones (generic, in-car, smart-home scripts), 1,796.70 hours from 3,442 speakers (ELRA catalogue ELRA-S0449)	512059.50 € (non-member, commercial use) per corpus of 1,796.70 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Top tier of a per-hour price ladder in this series (EUR 285 an hour). ELRA lists no non-commercial licence for it.	ELRA catalogue, German Speech Data by Mobile Phone - 1,796 Hours (ELRA-S0449) A catalogue or marketplace listing
German read speech recorded on mobile phones, 211 hours from 327 speakers (ELRA catalogue ELRA-S0473)	46103.50 € (non-member, commercial use) per corpus of 211 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, German Speech Data by Mobile Phone_Reading - 211 Hours (ELRA-S0473) A catalogue or marketplace listing
GlobalPhone Hausa: 8.73 hours of read Hausa speech (ELRA catalogue ELRA-S0347)	3600.00 € (non-member, commercial use) per corpus of 8.73 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. A small research corpus; per hour it is far dearer than larger commercial Hausa sets. A non-commercial-use licence is EUR 700 for an academic and EUR 3,600 for a company.	ELRA catalogue, GlobalPhone Hausa (ELRA-S0347) A catalogue or marketplace listing
GlobalPhone Tamil: 20 hours of read Tamil speech (ELRA catalogue ELRA-S0205)	600.00 € (non-member, commercial use) per corpus of 20 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Other GlobalPhone languages list at EUR 3,600 for about 20 hours, six times this price. A non-commercial-use licence is EUR 125 for an academic and EUR 600 for a company.	ELRA catalogue, GlobalPhone Tamil (ELRA-S0205) A catalogue or marketplace listing
Gram Vaani: 130 hours of Hindi phone recordings from 4,000 mostly rural speakers in Bihar, Jharkhand and Madhya Pradesh (ELRA catalogue ELRA-S0405)	50000.00 € (non-member, commercial use) per corpus of 130 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Rural, accented, real-world telephone speech from a community voice platform. A non-commercial-use licence is free for an academic and EUR 50,000 for a company.	ELRA catalogue, Gram Vaani data set (ELRA-S0405) A catalogue or marketplace listing
Greek SpeechDat(II) FDB-5000: 422 hours of fixed-network telephone speech from 5,000 speakers (ELRA catalogue ELRA-S0118)	70000.00 € (non-member, commercial use) per corpus of 422 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Telephone speech. A non-commercial-use licence is EUR 60,000 for an academic and EUR 70,000 for a company.	ELRA catalogue, Greek SpeechDat(II) FDB-5000 (ELRA-S0118) A catalogue or marketplace listing
Hindi prompted speech recorded on mobile phones, 759 hours from 1,425 speakers (ELRA catalogue ELRA-S0452)	115368.00 € (non-member, commercial use) per corpus of 759 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Bottom tier of the series' per-hour price ladder (EUR 152 an hour), shared with Indian English. ELRA lists no non-commercial licence for it.	ELRA catalogue, Hindi Speech Data by Mobile Phone - 759 Hours (ELRA-S0452) A catalogue or marketplace listing
Hindi read speech recorded on mobile phones, 240 hours from 401 speakers, transcribed (ELRA catalogue ELRA-S0463)	34200.00 € (non-member, commercial use) per corpus of 240 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, Hindi Speech Data by Mobile Phone_R - 240 Hours (ELRA-S0463) A catalogue or marketplace listing
Indonesian read speech recorded on mobile phones, 359 hours from 496 speakers (ELRA catalogue ELRA-S0470)	57978.50 € (non-member, commercial use) per corpus of 359 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, Indonesian Speech Data by Mobile Phone_R - 359 Hours (ELRA-S0470) A catalogue or marketplace listing
Japanese Kids Speech database (Lower Grade): 116.60 hours from 179 children aged 6 to 9, smartphone (ELRA catalogue ELRA-S0411)	12300.00 € (non-member, commercial use) per corpus of 116.60 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech, here priced below the adult Japanese sets. A non-commercial-use licence is EUR 2,100 for an academic and EUR 8,200 for a company.	ELRA catalogue, Japanese Kids Speech database (Lower Grade) (ELRA-S0411) A catalogue or marketplace listing
Japanese read speech recorded on mobile phones, 234.8 hours from 799 speakers (ELRA catalogue ELRA-S0476)	35689.60 € (non-member, commercial use) per corpus of 234.8 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, Japanese Speech Data by Mobile Phone_R - 234 Hours (ELRA-S0476) A catalogue or marketplace listing
Korean read speech recorded on mobile phones, 197 hours from 291 speakers (ELRA catalogue ELRA-S0475)	31815.50 € (non-member, commercial use) per corpus of 197 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, Korean Speech Data by Mobile Phone_Reading - 197 Hours (ELRA-S0475) A catalogue or marketplace listing
LILA Korean: 81.16 hours of mobile-network telephone speech from 1,000 speakers (ELRA catalogue ELRA-S0295)	51875.00 € (non-member, commercial use) per corpus of 81.16 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Mobile telephone speech in chosen acoustic conditions. A non-commercial-use licence is EUR 46,405 for an academic and EUR 51,875 for a company.	ELRA catalogue, LILA Korean database (ELRA-S0295) A catalogue or marketplace listing
MEDIA speech database: 70 hours of French human-machine dialogue (hotel booking scenario) (ELRA catalogue ELRA-S0272)	10000.00 € (non-member, commercial use) per corpus of 70 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Produced under a French government research programme. A non-commercial-use licence is free for an academic and EUR 10,000 for a company.	ELRA catalogue, MEDIA speech database for French (ELRA-S0272) A catalogue or marketplace listing
MoveOn Speech and Noise Corpus: 6 hours of British English command speech recorded on motorcycles (ELRA catalogue ELRA-S0370)	9000.00 € (non-member, commercial use) per corpus of 6 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Noisy speech for a niche domain (command speech on motorcycles, with a focus on police use); the dearest per hour of the ELRA prices recorded here. A non-commercial-use licence is EUR 1,500 for an academic and EUR 9,000 for a company.	ELRA catalogue, MoveOn Speech and Noise Corpus (ELRA-S0370) A catalogue or marketplace listing
OrienTel Egypt: colloquial Egyptian Arabic over fixed and mobile telephone networks, 50 hours from 750 speakers (ELRA catalogue ELRA-S0221)	60250.00 € (non-member, commercial use) per corpus of 50 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Egyptian Arabic here costs about 20 times SpeechData.ai's $65 an hour for Egyptian conversation. A non-commercial-use licence is EUR 45,187 for an academic and EUR 60,250 for a company.	ELRA catalogue, OrienTel Egypt MCA (Modern Colloquial Arabic) database (ELRA-S0221) A catalogue or marketplace listing
OrienTel Jordan: colloquial Arabic over fixed and mobile telephone networks, 50 hours from 757 speakers (ELRA catalogue ELRA-S0289)	60250.00 € (non-member, commercial use) per corpus of 50 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Dialect Arabic telephone speech at over EUR 1,200 an hour. A non-commercial-use licence is EUR 45,187 for an academic and EUR 60,250 for a company.	ELRA catalogue, OrienTel Jordan MCA (Modern Colloquial Arabic) database (ELRA-S0289) A catalogue or marketplace listing
OrienTel Tunisia: colloquial Arabic over fixed and mobile telephone networks, 50 hours from 792 speakers (ELRA catalogue ELRA-S0186)	27500.00 € (non-member, commercial use) per corpus of 50 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Dialect Arabic telephone speech. A non-commercial-use licence is EUR 16,500 for an academic and EUR 27,500 for a company.	ELRA catalogue, OrienTel Tunisia MCA (Modern Colloquial Arabic) database (ELRA-S0186) A catalogue or marketplace listing
Russian prompted speech recorded on mobile phones, 1,002 hours from 1,960 speakers (ELRA catalogue ELRA-S0443)	180861.00 € (non-member, commercial use) per corpus of 1,002 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.	ELRA catalogue, Russian Speech Data by Mobile Phone - 1,002 Hours (ELRA-S0443) A catalogue or marketplace listing
Spanish emotional speech synthesis database: 4 hours from two professional speakers in neutral and six emotional styles (ELRA catalogue ELRA-S0329)	4000.00 € (non-member, commercial use) per corpus of 4 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Emotional, acted speech for text-to-speech. A non-commercial-use licence is free for an academic and EUR 4,000 for a company.	ELRA catalogue, Emotional speech synthesis database (ELRA-S0329) A catalogue or marketplace listing
Spanish read speech recorded on mobile phones (Spain, Mexico, Venezuela), 227 hours, transcribed (ELRA catalogue ELRA-S0464)	40973.50 € (non-member, commercial use) per corpus of 227 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, Spanish Speech Data by Mobile Phone_R - 227 Hours (ELRA-S0464) A catalogue or marketplace listing
Castilian Spanish SpeechDat(II) FDB-4000: 246.67 hours of fixed-network telephone speech from 4,000 speakers (ELRA catalogue ELRA-S0102)	19500.00 € (non-member, commercial use) per corpus of 246.67 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Telephone speech. A non-commercial-use licence is EUR 12,500 for an academic and EUR 19,500 for a company.	ELRA catalogue, Spanish SpeechDat(II) FDB-4000 (ELRA-S0102) A catalogue or marketplace listing
Swiss-German SpeechDat(II) FDB-2000: 110 hours of fixed-network telephone speech from 2,000 speakers (ELRA catalogue ELRA-S0105)	38000.00 € (non-member, commercial use) per corpus of 110 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. A dialect with few corpora. A non-commercial-use licence is EUR 26,000 for an academic and EUR 38,000 for a company.	ELRA catalogue, Swiss-German SpeechDat(II) FDB-2000 (ELRA-S0105) A catalogue or marketplace listing
TC-STAR English training corpus: about 290 hours of European Parliament plenary speeches (ELRA catalogue ELRA-S0251)	9000.00 € (non-member, commercial use) per corpus of 290 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Public parliamentary speech, priced well below collected speech. A non-commercial-use licence is EUR 6,270 for an academic and EUR 9,000 for a company.	ELRA catalogue, TC-STAR English Training Corpora for ASR: Recordings of EPPS Speech (ELRA-S0251) A catalogue or marketplace listing
TC-STAR Spanish baseline female voice: 10 hours of one speaker reading in a noise-reduced room, for speech synthesis (ELRA catalogue ELRA-S0309)	7800.00 € (non-member, commercial use) per corpus of 10 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Single-speaker studio voice data for text-to-speech. A non-commercial-use licence is EUR 3,850 for an academic and EUR 7,800 for a company.	ELRA catalogue, TC-STAR Spanish Baseline Female Speech Database (ELRA-S0309) A catalogue or marketplace listing
Thai read speech recorded on mobile phones, 203 hours from 498 speakers (ELRA catalogue ELRA-S0471)	47093.40 € (non-member, commercial use) per corpus of 203 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.	ELRA catalogue, Thai Speech Data by Mobile Phone_Reading - 203 Hours (ELRA-S0471) A catalogue or marketplace listing
THERADIA WoZ corpus: 86 hours of audiovisual sessions for detecting affect in home cognitive-training (healthcare) (ELRA catalogue ELRA-S0499)	6000.00 € (non-member, commercial use) per corpus of 86 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Healthcare-domain affective speech. A non-commercial-use licence is free for an academic and EUR 6,000 for a company.	ELRA catalogue, THERADIA WoZ Corpus (ELRA-S0499) A catalogue or marketplace listing
TRAD Pashto Broadcast News: 108 hours of transcribed TV and radio news from more than 1,000 speakers (ELRA catalogue ELRA-S0381)	28000.00 € (non-member, commercial use) per corpus of 108 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. A low-resource language, priced well above broadcast speech in big languages. A non-commercial-use licence is EUR 3,500 for an academic and EUR 28,000 for a company.	ELRA catalogue, TRAD Pashto Broadcast News Speech Corpus (ELRA-S0381) A catalogue or marketplace listing
UK English Speecon: 204 hours from 606 adults and 51 children, 4 microphone channels in office, car and public places (ELRA catalogue ELRA-S0215)	75000.00 € (non-member, commercial use) per corpus of 204 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Every Speecon corpus seen in this catalogue lists at EUR 75,000, whatever the language. A non-commercial-use licence is EUR 60,000 for an academic and EUR 75,000 for a company.	ELRA catalogue, UK English Speecon database (ELRA-S0215) A catalogue or marketplace listing
Vietnamese prompted speech recorded on mobile phones, 760 hours from 1,751 speakers (ELRA catalogue ELRA-S0432)	122740.00 € (non-member, commercial use) per corpus of 760 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.	ELRA catalogue, Vietnamese Speech Data by Mobile Phone - 760 Hours (ELRA-S0432) A catalogue or marketplace listing
Welsh SpeechDat(II) FDB-2000: 170 hours of Welsh telephone speech (ELRA catalogue ELRA-S0075)	18000.00 € (non-member, commercial use) per corpus of 170 hours	Oct 2026	ELRA's non-member price for a company under the commercial-use (VAR) licence. A minority European language. A non-commercial-use licence is EUR 11,000 for an academic and EUR 18,000 for a company.	ELRA catalogue, Welsh SpeechDat(II) FDB-2000 (ELRA-S0075) A catalogue or marketplace listing
AIDA Ukrainian Broadcast and Telephone Speech: about 156 hours with 1.2 million words of transcripts	$2,000.00 (non-member fee) per corpus of about 156 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.	LDC catalog, AIDA Ukrainian Broadcast and Telephone Speech Audio and Transcripts (LDC2023S01) A catalogue or marketplace listing
AISHELL-1: about 520 hours of Mandarin read speech from 400 speakers on three devices	$3,000.00 (non-member fee) per corpus of about 520 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, AISHELL-1 (LDC2018S14) A catalogue or marketplace listing
AnnoDIFP Session Audio and Transcripts: 438.34 hours of English in-person interview audio with personality scores	$3,000.00 (non-member fee) per corpus of about 438.34 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, AnnoDIFP Session Audio and Transcripts (LDC2025S06) A catalogue or marketplace listing
Air Traffic Control Complete: about 70 hours of controller and pilot radio transmissions	$1,150.00 (non-member fee) per corpus of about 70 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $575.00.	LDC catalog, Air Traffic Control Complete (LDC94S14A) A catalogue or marketplace listing
BOLT CTS CALLFRIEND CALLHOME Egyptian Arabic Audio: about 116 hours of telephone conversations	$2,000.00 (non-member fee) per corpus of about 116 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.	LDC catalog, BOLT CTS CALLFRIEND CALLHOME Egyptian Arabic Audio (LDC2025S09) A catalogue or marketplace listing
Call My Net 1: 364 hours of conversational telephone speech in Tagalog, Cebuano, Cantonese and Mandarin	$2,500.00 (non-member fee) per corpus of about 364 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.	LDC catalog, Call My Net 1 (LDC2024S05) A catalogue or marketplace listing
CALLHOME American English Second Edition: about 56 hours of unscripted telephone conversations	$3,000.00 (non-member fee) per corpus of about 56 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.	LDC catalog, CALLHOME American English Second Edition (LDC2026S08) A catalogue or marketplace listing
CALLHOME German Second Edition: about 48 hours of unscripted telephone conversations	$3,000.00 (non-member fee) per corpus of about 48 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.	LDC catalog, CALLHOME German Second Edition (LDC2026S06) A catalogue or marketplace listing
CALLHOME Japanese Second Edition: about 49 hours of unscripted telephone conversations	$3,000.00 (non-member fee) per corpus of about 49 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, CALLHOME Japanese Second Edition (LDC2026S02) A catalogue or marketplace listing
CALLHOME Mandarin Chinese Second Edition: about 38 hours of unscripted telephone conversations	$3,000.00 (non-member fee) per corpus of about 38 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.	LDC catalog, CALLHOME Mandarin Chinese Second Edition (LDC2026S11) A catalogue or marketplace listing
CALLHOME Spanish Second Edition: about 38 hours of unscripted telephone conversations	$3,000.00 (non-member fee) per corpus of about 38 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, CALLHOME Spanish Second Edition (LDC2026S04) A catalogue or marketplace listing
Fisher English Training Speech Part 1: 984 hours of English conversational telephone speech	$7,000.00 (non-member fee) per corpus of about 984 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $3,500.00.	LDC catalog, Fisher English Training Speech Part 1 Speech (LDC2004S13) A catalogue or marketplace listing
Fisher Levantine Arabic Conversational Telephone Speech: 45 hours (279 calls)	$1,000.00 (non-member fee) per corpus of about 45 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $500.00.	LDC catalog, Fisher Levantine Arabic Conversational Telephone Speech (LDC2007S02) A catalogue or marketplace listing
Fisher Spanish Speech: roughly 163 hours of Spanish telephone speech from 136 speakers	$2,500.00 (non-member fee) per corpus of about 163 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.	LDC catalog, Fisher Spanish Speech (LDC2010S01) A catalogue or marketplace listing
GALE Phase 2 Arabic Broadcast News Speech Part 1: about 165 hours of Arabic broadcast news	$2,000.00 (non-member fee) per corpus of about 165 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.	LDC catalog, GALE Phase 2 Arabic Broadcast News Speech Part 1 (LDC2014S07) A catalogue or marketplace listing
Global TIMIT Learner Treebank English: about 24 hours of native and non-native (L2) English read speech	$750.00 (non-member fee) per corpus of about 24 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $375.00.	LDC catalog, Global TIMIT Learner Treebank English (LDC2020S09) A catalogue or marketplace listing
HKUST Mandarin Telephone Speech Part 1: about 149 hours of Mandarin conversational telephone speech	$3,000.00 (non-member fee) per corpus of about 149 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, HKUST Mandarin Telephone Speech, Part 1 (LDC2005S15) A catalogue or marketplace listing
IARPA Babel Zulu Language Pack: about 211 hours of Zulu conversational and scripted telephone speech	$25.00 (non-member fee) per corpus of about 211 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $25.00.	LDC catalog, IARPA Babel Zulu Language Pack IARPA-babel206b-v0.1e (LDC2017S19) A catalogue or marketplace listing
ICSI Meeting Speech: about 72 hours of English meeting speech (883 hours of multi-microphone audio)	$1,900.00 (non-member fee) per corpus of about 72 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $950.00.	LDC catalog, ICSI Meeting Speech (LDC2004S02) A catalogue or marketplace listing
Korean Telephone Conversations Speech: about 44 hours of audio (100 calls)	$1,500.00 (non-member fee) per corpus of about 44 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $750.00.	LDC catalog, Korean Telephone Conversations Speech (LDC2003S03) A catalogue or marketplace listing
King Saud University Arabic Speech Database: 590 hours of Arabic speech from 269 speakers	$2,000.00 (non-member fee) per corpus of about 590 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.	LDC catalog, King Saud University Arabic Speech Database (LDC2014S02) A catalogue or marketplace listing
Magic Data Chinese Mandarin Conversational Speech: about 10 hours of conversation from 60 speakers (about 60 hours of multi-device audio)	$750.00 (non-member fee) per corpus of about 10 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $375.00.	LDC catalog, Magic Data Chinese Mandarin Conversational Speech (LDC2019S23) A catalogue or marketplace listing
MATERIAL Somali-English Language Pack: about 80 hours of Somali telephone speech with transcripts and translations	$250.00 (non-member fee) per corpus of about 80 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $125.00.	LDC catalog, MATERIAL Somali-English Language Pack (LDC2024S10) A catalogue or marketplace listing
MATERIAL Swahili-English Language Pack: about 112 hours of Swahili telephone speech with transcripts and translations	$250.00 (non-member fee) per corpus of about 112 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.	LDC catalog, MATERIAL Swahili-English Language Pack (LDC2026S01) A catalogue or marketplace listing
Mixer 6 CHiME 8 Transcribed Calls and Interviews: 80 hours of English interviews and telephone speech with transcripts	$3,000.00 (non-member fee) per corpus of about 80 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, Mixer 6 - CHiME 8 Transcribed Calls and Interviews (LDC2025S07) A catalogue or marketplace listing
Multi-Language Conversational Telephone Speech 2011, South Asian: about 118 hours of Bengali, Hindi, Punjabi, Tamil and Urdu telephone speech	$2,500.00 (non-member fee) per corpus of about 118 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.	LDC catalog, Multi-Language Conversational Telephone Speech 2011 -- South Asian (LDC2017S14) A catalogue or marketplace listing
Multi-Channel WSJ Audio: about 100 hours of British English read speech from 45 speakers, far-field microphone arrays	$1,500.00 (non-member fee) per corpus of about 100 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $750.00.	LDC catalog, Multi-Channel WSJ Audio (LDC2014S03) A catalogue or marketplace listing
Polish Speech Database: about 280 hours of read Polish speech from 200 speakers	$3,000.00 (non-member fee) per corpus of about 280 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, Polish Speech Database (LDC2019S19) A catalogue or marketplace listing
REMIX Telephone Collection: 320 hours of English conversational telephone speech from 358 speakers	$2,500.00 (non-member fee) per corpus of about 320 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.	LDC catalog, REMIX Telephone Collection (LDC2023S09) A catalogue or marketplace listing
Samromur Children Icelandic Speech 1.0: 131 hours of prompted speech from 3,175 children aged 4 to 17	$250.00 (non-member fee) per corpus of about 131 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $125.00.	LDC catalog, Samrómur Children Icelandic Speech 1.0 (LDC2022S11) A catalogue or marketplace listing
Mandarin-English Code-Switching in South-East Asia (SEAME): about 192 hours from 156 speakers	$2,000.00 (non-member fee) per corpus of about 192 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.	LDC catalog, Mandarin-English Code-Switching in South-East Asia (LDC2015S04) A catalogue or marketplace listing
Switchboard-1 Release 2: about 260 hours of US English conversational telephone speech (543 speakers)	$3,000.00 (non-member fee) per corpus of about 260 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.	LDC catalog, Switchboard-1 Release 2 (LDC97S62) A catalogue or marketplace listing
TORGO Database of Dysarthric Articulation: about 23 hours of English speech from speakers with cerebral palsy or ALS and controls	$1,200.00 (non-member fee) per corpus of about 23 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $600.00.	LDC catalog, TORGO Database of Dysarthric Articulation (LDC2012S02) A catalogue or marketplace listing
United Nations Proceedings Speech: about 8,500 hours of proceedings in the six official UN languages	$5,000.00 (non-member fee) per corpus of about 8,500 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $2,500.00.	LDC catalog, United Nations Proceedings Speech (LDC2014S08) A catalogue or marketplace listing
CSR-I (WSJ0) Complete: about 141 hours of English read speech (123 speakers reading the Wall Street Journal)	$1,500.00 (non-member fee) per corpus of about 141 hours	Oct 2026	LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $750.00.	LDC catalog, CSR-I (WSJ0) Complete (LDC93S6A) A catalogue or marketplace listing
35,000-hour multilingual conversational speech dataset (Nexdata, 100+ languages)	$20,000 / purchase (starts at) per purchase (one-off)	Oct 2026	A starting list price for bulk off-the-shelf audio, well under $1 per hour if the whole set is bought; final price is quoted per scope.	Datarade (Nexdata listing) A vendor's published price guide
Arabic podcast audio, dual-channel, 6,024 hours (Infobay.AI on Opendatabay)	$29,300 per dataset of 6,024 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-16.	Opendatabay marketplace, 6K+ Hours Arabic Dual-Channel Podcast Audio Dataset A catalogue or marketplace listing
English podcast audio, dual-channel, 1,165 hours (Infobay.AI on Opendatabay)	$5,660 per dataset of 1,165 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-17.	Opendatabay marketplace, 1K+ Hours English Dual-Channel Podcast Audio Dataset A catalogue or marketplace listing
German call-center recordings, dual-channel, 212 hours (Infobay.AI on Opendatabay)	$1,320 per dataset of 212 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-11.	Opendatabay marketplace, 200+ Hours German Dual-Channel Call Center Dataset A catalogue or marketplace listing
Hindi call-center recordings, dual-channel, 1,587,658 hours (Infobay.AI on Opendatabay)	$826,700 per dataset of 1,587,658 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-11.	Opendatabay marketplace, 1.5Mn+ Hours Hindi Dual-Channel Call Center Dataset A catalogue or marketplace listing
Hindi podcast audio, dual-channel, 11,607 hours (Infobay.AI on Opendatabay)	$56,400 per dataset of 11,607 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-17.	Opendatabay marketplace, 11K+ Hours Hindi Dual-Channel Podcast Audio Dataset A catalogue or marketplace listing
Mizo (low-resource language of north-east India) call-center recordings, dual-channel, 468 hours (Infobay.AI on Opendatabay)	$2,370 per dataset of 468 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.	Opendatabay marketplace, 400+ Hours Mizo Dual-Channel Call Center Dataset A catalogue or marketplace listing
Nepali call-center recordings, dual-channel, 229,645 hours (Infobay.AI on Opendatabay)	$487,100 per dataset of 229,645 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.	Opendatabay marketplace, 229K+ Hours Nepali Dual-Channel Call Center Dataset A catalogue or marketplace listing
Odia call-center recordings, dual-channel, 12,794 hours (Infobay.AI on Opendatabay)	$62,400 per dataset of 12,794 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.	Opendatabay marketplace, 12K+ Hours Odia Dual-Channel Call Center Dataset A catalogue or marketplace listing
Somali call-center recordings, dual-channel, 1,057 hours (Infobay.AI on Opendatabay)	$5,260 per dataset of 1,057 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.	Opendatabay marketplace, 1K+ Hours Somali Dual-Channel Call Center Dataset A catalogue or marketplace listing
Swahili call-center recordings, dual-channel, 194,331 hours (Infobay.AI on Opendatabay)	$292,100 per dataset of 194,331 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.	Opendatabay marketplace, 194K+ Hours Swahili Dual-Channel Call Center Dataset A catalogue or marketplace listing
Tamil call-center recordings, dual-channel, 15,056 hours (Infobay.AI on Opendatabay)	$73,300 per dataset of 15,056 hours	Oct 2026	Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.	Opendatabay marketplace, 15K+ Hours Tamil Dual-Channel Call Center Dataset A catalogue or marketplace listing
Sound-effects catalogue for AI training: 350+ libraries, about 53,000 files (vehicles, machinery, ambiences, foley) from Krampfstadt Studio, on Opendatabay	$78,900 per catalogue of 53k sound files	Oct 2026	A studio's whole sound-effects back catalogue offered for AI training. The same seller lists it on Datarade at EUR 60,000 one-off, so the asking price differs by marketplace.	Opendatabay marketplace, Krampfstadt Studio 350+ SFX libraries A catalogue or marketplace listing
Saudi Arabic full-duplex customer-service dialogue speech (268 speakers, transcribed), 268 hours (Nexdata on Opendatabay)	$118,400 per dataset of 268 hours	Oct 2026	Nexdata's asking price on the Opendatabay marketplace for a commercial licence with instant download; Gulf Arabic customer-service dialogue is among the dearest speech on the market. Listed 2026-08-14.	Opendatabay marketplace, 268 Hours Arabic Saudi Full Duplex Speech Data A catalogue or marketplace listing
INTERSPEECH 2025 MLC-SLM challenge data: conversational speech in 11 languages, 1,500 hours, curated by Datatang (Nexdata on Opendatabay)	$526,300 per dataset of 1,500 hours	Oct 2026	Nexdata's asking price on Opendatabay for the commercial licence of a set first released for a research challenge; Nexdata is Datatang's international brand. Listed 2026-08-14.	Opendatabay marketplace, INTERSPEECH 2025 MLC SLM Challenge Data A catalogue or marketplace listing
Tagalog full-duplex multi-channel spontaneous dialogue speech (smartphone, transcribed), 1,100 hours (Nexdata on Opendatabay)	$394,700 per dataset of 1,100 hours	Oct 2026	Nexdata's asking price on the Opendatabay marketplace for a commercial licence with instant download. Full-duplex, speaker-separated dialogue is the premium tier of conversational speech. Listed 2026-08-14.	Opendatabay marketplace, 1100 Hours Tagalog Full Duplex Multi Channel Speech Data A catalogue or marketplace listing
US English real customer-service calls, dual-channel, with verbatim transcripts, 1,000 hours (PissedConsumer on Opendatabay)	$24,500 per dataset of 1,000 hours	Oct 2026	Asking price for a 1,000-hour package of real (not staged) PII-redacted support calls; larger volumes are quoted on request. Listed 2026-07-30.	Opendatabay marketplace, Audio Data: English (US) Conversational Speech Audio Dataset A catalogue or marketplace listing
Hindi real customer-service calls with Hindi-English code-switching, dual-channel, with verbatim transcripts, 1,000 hours (PissedConsumer on Opendatabay)	$24,500 per dataset of 1,000 hours	Oct 2026	Same seller and same price as its US English package, so this seller does not charge a language premium for Hindi. Listed 2026-07-30.	Opendatabay marketplace, Audio Data: Hindi Conversational Speech Audio Dataset A catalogue or marketplace listing
Human transcription of English audio, 99%+ accuracy, 12-hour turnaround (Rev)	$1.99 /min per audio minute	Oct 2026	Pay-as-you-go list price for human transcription, the labelling step that turns raw audio into ASR training data; subscriptions save up to 15%.	Rev pricing page A vendor's published price guide
Amharic (Ethiopia) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Amharic (Ethiopia) conversational speech dataset A catalogue or marketplace listing
Arabic (Egyptian) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Arabic (Egyptian) conversational speech dataset A catalogue or marketplace listing
Arabic (Modern Standard, Saudi Arabia) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$90 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Arabic (Modern Standard, Saudi Arabia) conversational speech dataset A catalogue or marketplace listing
Bengali (Bangladesh) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Bengali (Bangladesh) conversational speech dataset A catalogue or marketplace listing
English (South African) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$60 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, English (South African) conversational speech dataset A catalogue or marketplace listing
English (US) conversational speech dataset, 1,500 hours of two-person conversations from 150 native speakers, transcribed, commercial licence	$90 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, English (US) conversational speech dataset A catalogue or marketplace listing
French (DR Congo) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, French (DR Congo) conversational speech dataset A catalogue or marketplace listing
French (France) conversational speech dataset, 1,000 hours of two-person conversations from 100 native speakers, transcribed, commercial licence	$95 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, French (France) conversational speech dataset A catalogue or marketplace listing
German conversational speech dataset, 1,000 hours of two-person conversations from 50 native speakers, transcribed, commercial licence	$95 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, German conversational speech dataset A catalogue or marketplace listing
Hausa (Nigeria) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Hausa (Nigeria) conversational speech dataset A catalogue or marketplace listing
Hindi conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Hindi conversational speech dataset A catalogue or marketplace listing
Japanese conversational speech dataset, 500 hours of two-person conversations from 50 native speakers, transcribed, commercial licence	$95 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Japanese conversational speech dataset A catalogue or marketplace listing
Korean conversational speech dataset, 500 hours of two-person conversations from 50 native speakers, transcribed, commercial licence	$95 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Korean conversational speech dataset A catalogue or marketplace listing
Mandarin Chinese (Taiwan) conversational speech dataset, 1,500 hours of two-person conversations from 150 native speakers, transcribed, commercial licence	$70 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Mandarin Chinese (Taiwan) conversational speech dataset A catalogue or marketplace listing
Portuguese (Brazil) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Portuguese (Brazil) conversational speech dataset A catalogue or marketplace listing
Spanish (Mexico) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Spanish (Mexico) conversational speech dataset A catalogue or marketplace listing
Spanish (Spain) conversational speech dataset, 1,000 hours of two-person conversations from 100 native speakers, transcribed, commercial licence	$85 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Spanish (Spain) conversational speech dataset A catalogue or marketplace listing
Swahili (Kenya) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Swahili (Kenya) conversational speech dataset A catalogue or marketplace listing
Tamil conversational speech dataset, 500 hours of two-person conversations from 50 native speakers, transcribed, commercial licence	$75 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Tamil conversational speech dataset A catalogue or marketplace listing
Urdu (Pakistan) conversational speech dataset, 1,000 hours of two-person conversations from 100 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Urdu (Pakistan) conversational speech dataset A catalogue or marketplace listing
Yoruba (Nigeria) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence	$65 per hour of audio per hour of audio	Oct 2026	Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.	SpeechData.ai catalogue, Yoruba (Nigeria) conversational speech dataset A catalogue or marketplace listing
Human transcription of audio and video, 24-hour turnaround (Speechpad)	$1.30 per minute per audio minute	Oct 2026	List price for standard human transcription; 6 to 12 hour turnaround is $1.80 a minute, and timestamps add $0.25 and verbatim $0.35 a minute, the two features ASR training data needs.	Speechpad pricing page A vendor's published price guide
Human-edited transcription, 99%+ accuracy, 1 to 3 day turnaround (TranscribeMe)	Starting at $0.79/min per audio minute (starting price)	Oct 2026	A starting price, so it marks the floor of human transcription at a vendor that also sells AI training data services.	TranscribeMe pricing page A vendor's published price guide
British English recordings evaluated for pronunciation, fluency, tone and audio quality	$40 to $65 per hour per hour worked (paid to the evaluator)	Oct 2026	Evaluating speech pays more than producing or transcribing it on the same board. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, English-Speaking (British) Audio Evaluator The payer's own page
Spanish (Mexico) voice actor recording scripted and emotional audio for AI training	$20 to $65 per hour per hour worked (paid to the voice actor)	Oct 2026	Hourly pay for voice acting for AI; the job board is the hiring organisation for its client. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Spanish-Speaking Voice Actor for AI Training The payer's own page
Bengali audio and video transcription and machine-transcript review	$10 to $22 per hour per hour worked (paid to the transcriber)	Oct 2026	Transcriber pay for Bengali. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Bengali-Speaking Audio and Video Transcription Expert The payer's own page
Gujarati audio transcription and review for AI training	$8 to $12 per hour per hour worked (paid to the transcriber)	Oct 2026	Transcriber pay for an Indic language, the low end of transcription pay on the board. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Gujarati-Speaking Audio Transcription Expert The payer's own page
Norwegian audio and video transcription and speech-segment classification	$20 to $30 per hour per hour worked (paid to the transcriber)	Oct 2026	Transcriber pay for a small, high-wage European language. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Norwegian-Speaking Audio Transcription Expert The payer's own page
Punjabi audio and video transcription with speaker identification and metadata	$8 to $15 per hour per hour worked (paid to the transcriber)	Oct 2026	Transcriber pay for an Indic language. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Punjabi-Speaking Transcription Expert The payer's own page
Voice owner's royalty when others generate speech with their cloned voice in the ElevenLabs Voice Library (default rate)	$0.03 per 1,000 characters per 1,000 characters generated (paid to the voice owner)	Oct 2026	A per-use royalty to the person whose voice is cloned, paid weekly; it counts paid users' generations only. The page says voice actors have earned a combined $5 million through the library.	ElevenLabs blog, How to monetize your voice with the Voice Library The payer's own page
Voice owner's royalty in the ElevenLabs Voice Library for HQ or rare voices (maximum requestable rate)	up to $0.20 per 1,000 characters per 1,000 characters generated (paid to the voice owner)	Oct 2026	The ceiling a voice owner can set; a rare accent or tone is what earns it.	ElevenLabs blog, How to monetize your voice with the Voice Library The payer's own page
Expressive voice recording for AI training, Bengali speakers in India (Outlier, via Braintrust)	up to $7.50 USD per hour per hour, ceiling for core project work (paid to the speaker)	Oct 2026	The same Outlier voice project pays by the speaker's language and country: $7.50 for Bengali in India, $10 Thai, $11 Vietnamese, $20 Simplified Chinese, $25 French. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Bengali Expressive Voice AI Trainer at Outlier via Braintrust A catalogue or marketplace listing
Expressive voice recording for AI training, Simplified Chinese speakers (Outlier, via Braintrust)	up to $20 USD per hour per hour, ceiling for core project work (paid to the speaker)	Oct 2026	One of a set of identical Outlier voice projects priced by language; open in Australia, Canada, the UK, Malaysia, New Zealand and Singapore. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Expressive Voice AI Trainer - Simplified Chinese at Outlier via Braintrust A catalogue or marketplace listing
Expressive voice recording for AI training, French speakers (Outlier, via Braintrust)	up to $25 USD per hour per hour, ceiling for core project work (paid to the speaker)	Oct 2026	The top rate in a set of identical Outlier voice projects priced by language. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, French Expressive Voice AI Trainer at Outlier via Braintrust A catalogue or marketplace listing
Expressive voice recording for AI training, Thai speakers (Outlier, via Braintrust)	up to $10 USD per hour per hour, ceiling for core project work (paid to the speaker)	Oct 2026	One of a set of identical Outlier voice projects priced by language. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Thai Expressive Voice AI Trainer at Outlier via Braintrust A catalogue or marketplace listing
Expressive voice recording for AI training, Vietnamese speakers in Vietnam (Outlier, via Braintrust)	up to $11 USD per hour per hour, ceiling for core project work (paid to the speaker)	Oct 2026	One of a set of identical Outlier voice projects priced by language. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Vietnamese Expressive Voice AI Trainer at Outlier via Braintrust A catalogue or marketplace listing
Spanish voice-over recording for AI training and media, by finished hour of audio (professional home studio)	USD 200 per finished hour per finished hour of recorded audio (paid to the voice talent)	Sep 2026	Pay per delivered hour of edited, mastered audio, not per hour worked; professional studio-quality equipment required. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Spanish-Speaking Voice Recording Specialist The payer's own page
Japanese audio transcription, error correction and context notes for AI training	$20 to $36 per hour per hour worked (paid to the transcriber)	Sep 2026	Transcriber pay for Japanese, roughly double the Indic-language rates on the same board. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Japanese-Speaking Audio Transcription Expert The payer's own page
English phone calls to US businesses (reservations, appointments) made by Latin American contractors to help train AI systems	$18 to $28 per hour per scheduled hour (paid to the caller)	Sep 2026	Paid for scheduled availability in four 5-hour shifts a week, open to Argentina, Brazil, Chile, Colombia, Mexico and Peru. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, English-Speaking Customer Service Phone Support Expert The payer's own page
Mandarin and English audio transcription and review for AI training	$20 to $30 per hour equivalent per hour equivalent, paid per task (paid to the transcriber)	Sep 2026	Paid per task; the hourly figure is the job's own equivalent. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Mandarin-Speaking Audio Transcription Expert The payer's own page
Portuguese audio transcription and review, with occasional English	$10 to $18 per hour equivalent per hour equivalent, paid per task (paid to the transcriber)	Sep 2026	Paid per task; the hourly figure is the job's own equivalent. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Portuguese-Speaking Audio Transcription Expert The payer's own page
Bilingual Hebrew and English audio contributor for AI training (AfterQuery)	$30/hr per hour (paid to the contributor)	Sep 2026	Posted about a month before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Bilingual Audio Contributor (Hebrew/English) at AfterQuery A catalogue or marketplace listing
Paid English conversation recording study for AI training (RemoExperts, open worldwide)	$20/hr per hour (paid to the speaker)	Sep 2026	Posted about two weeks before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Paid English Conversation Recording Study at RemoExperts A catalogue or marketplace listing
Tamil audio transcription and review, varied accents and dialects	$10 to $24 per hour equivalent per hour equivalent, paid per task (paid to the transcriber)	Aug 2026	Paid per task; the hourly figure is the job's own equivalent. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Tamil-Speaking Audio Transcription Expert The payer's own page
Mandarin-English code-switched conversation recording (1 to 2 minute clips) by speakers in Singapore or Malaysia	$2 to $5 USD per hour per hour worked (paid to the speaker)	Aug 2026	The lowest speaker pay found in this run, for phone-recorded clips with no studio needed. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, Chinese- and English-Speaking Bilingual Audio Recording The payer's own page
German speech data: recording audio samples, phonetic annotation and evaluation	$30 to $65 per hour per hour worked (paid to the contributor)	Aug 2026	Mixes recording with phonetic annotation and review, so it prices a skilled native speaker rather than a voice alone. OpenTrain states it is the hiring organisation for the client project.	OpenTrain AI job posting, German-Speaking Speech Data Annotation Expert The payer's own page
Voice actor recording a customer-service agent voice for cloning, African American English (Mercor)	$50–$100/hr per hour (paid to the voice actor)	Aug 2026	Posted about two months before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Voice Actor: CX Agent Voice Cloning (African American) at Mercor A catalogue or marketplace listing
Voice actor recording a customer-service agent voice for cloning, Swiss German (Mercor)	$50–$150/hr per hour (paid to the voice actor)	Aug 2026	Posted about two months before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Voice Actor: CX Agent Voice Cloning (Swiss German) at Mercor A catalogue or marketplace listing
Voice recorder, German and English (micro1)	$30–$65/hr per hour (paid to the speaker)	Aug 2026	Posted about two months before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Voice Recorder (German + English) at micro1 A catalogue or marketplace listing
Audio recording contributor with an Australian English accent (AfterQuery)	$50/hr per hour (paid to the speaker)	Jul 2026	A specific accent commands more than general English recording at the same buyer. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Audio Recording Contributor - Australian English Accent at AfterQuery A catalogue or marketplace listing
French conversation partner recording natural conversations for AI training (Babel Audio, Canada and US)	$17.50/hr per hour (paid to the speaker)	Jul 2026	Posted about three months before it was read; Babel Audio pays per recorded hour. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, French Conversation Partner (AI Training) at Babel Audio A catalogue or marketplace listing
Royalties paid to voice creators for licensing their cloned voices on the ElevenLabs Voice Library, cumulative to May 2026	over $22 million in total, to 10,400+ creators	May 2026	Royalties for the use of a voice, not a fee for training data; up from $11 million in November 2025. Over $22 million across 10,400+ creators is about $2,100 per creator on average (our arithmetic).	ElevenLabs blog, $22 million earned by voice creators on ElevenLabs The payer's own page
Hindi voice actor or actress recording for AI training (Babel Audio)	$60–$150/hr per hour (paid to the voice actor)	May 2026	Professional voice acting pays far above conversational recording at the same platform ($17.50 an hour). Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Hindi Voice Actor / Actress at Babel Audio A catalogue or marketplace listing
Conversation recording roles on David AI's Babel Audio platform	$24 to $28 per hour per hour (paid to the speaker)	May 2026	Speaker-side pay at an audio dataset vendor that has closed six-figure contracts with large AI labs; the buyer price is not public.	Contrary Research Press report
Voice actor recording a customer-service agent voice for cloning, US English (Mercor)	$50–$150/hr per hour (paid to the voice actor)	Apr 2026	Voice-cloning work, where the actor's own voice becomes the product, pays the most of any speaker role found. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.	AITraining.jobs, Voice Actor: CX Agent Voice Cloning (USA) at Mercor A catalogue or marketplace listing
Speech collected to a buyer's spec	$150 to $400+ an hour per hour	Jan 2026	What a buyer pays a vendor to record new speech: chosen speakers, languages and scenarios, with consent on file.	SpeechData.ai buyer's guide A vendor's published price guide
Recorded phone-call audio sold to AI developers (Neon Mobile app, paid to callers)	$0.15 per minute, doubled when both callers use the app, max $30 per day per minute (paid to the caller)	Sep 2025	This is what the app pays the person recording, not what AI buyers pay Neon; the app later went offline after a security flaw.	Malwarebytes Press report
Open African-language speech collection: 9,000 hours across 18 languages in Kenya, Nigeria and South Africa (Gates Foundation grant)	$2.2 million per programme of 9,000 hours	Sep 2025	What a funded collection programme cost per hour, including recruiting speakers, recording, transcription and research staff; the data was released open access, so no seller recovers it.	CO/AI news, African researchers release 9,000-hour speech dataset Press report
Defined.ai Data Access Plan, Elite tier: 3,125 hours of off-the-shelf speech data from its marketplace	€500,000 for 3,125 hours per plan of 3,125 hours	Sep 2024	A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.	Defined.ai press release, Data Access Plan (DAP) The payer's own page
Defined.ai Data Access Plan, Premium tier: 625 hours of off-the-shelf speech data from its marketplace	€100,000 for 625 hours per plan of 625 hours	Sep 2024	A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.	Defined.ai press release, Data Access Plan (DAP) The payer's own page
Defined.ai Data Access Plan, Standard tier: 315 hours of off-the-shelf speech data from its marketplace	€50,000 for 315 hours per plan of 315 hours	Sep 2024	A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.	Defined.ai press release, Data Access Plan (DAP) The payer's own page
Defined.ai Data Access Plan, Ultimate tier: 1,565 hours of off-the-shelf speech data from its marketplace	€250,000 for 1,565 hours per plan of 1,565 hours	Sep 2024	A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.	Defined.ai press release, Data Access Plan (DAP) The payer's own page
Rural Indian speakers recording speech data for Karya (supplier to Microsoft and Google)	about $5 an hour per hour (paid to the speaker)	Dec 2023	Worker pay, well above the Indian minimum wage; workers also keep a share of the data for royalties.	Context (Thomson Reuters Foundation) Press report
One hour of Odia (low-resource Indian language) speech data	used to cost about $3-$4, now it's $40 per hour of audio	Dec 2023	Karya co-founder Safiya Husain's figure for what buyers now pay; a tenfold rise as big tech sought Indian-language data.	Context (Thomson Reuters Foundation) Press report
Corpus of Spontaneous Japanese (CSJ), 661 hours of mostly monologue speech with 7.52 million words of transcripts: a 2-year commercial licence for up to 10 users	500,000 yen (tax not included) per licence of the 661-hour corpus	Jun 2021	The corpus owner's own commercial fee schedule (last updated 1 June 2021); an unlimited-user commercial licence is 10 million yen. The same corpus costs EUR 43,125 through ELRA.	NINJAL, Corpus of Spontaneous Japanese usage fees A catalogue or marketplace listing
Corpus of Spontaneous Japanese (CSJ), 661 hours of mostly monologue speech with 7.52 million words of transcripts: an open-ended commercial licence for up to 10 users	5,000,000 yen (tax not included) per licence of the 661-hour corpus	Jun 2021	The corpus owner's own commercial fee schedule (last updated 1 June 2021); an unlimited-user commercial licence is 10 million yen. The same corpus costs EUR 43,125 through ELRA.	NINJAL, Corpus of Spontaneous Japanese usage fees A catalogue or marketplace listing
Off-the-shelf speech recognition training datasets (wake words, read sentences, dialogues; Chinese and foreign languages), usage licences sold by Haitian Ruisheng in 2020: 198,951 hours	413 元/小时 (CNY 413 an hour, average over 198,951 hours sold) per hour, average across all speech recognition dataset hours licensed in 2020	Dec 2020	Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2020, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate. The same filing splits 2020 by language: Chinese 91,455 hours at CNY 255 an hour, foreign languages 108,525 hours at CNY 674 an hour (all speech, both kinds).	Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table The payer's own page
Off-the-shelf speech synthesis (TTS) training datasets, usage licences sold by Haitian Ruisheng in 2020: 1,029 hours	13,832 元/小时 (CNY 13,832 an hour, average over 1,029 hours sold) per hour, average across all speech synthesis dataset hours licensed in 2020	Dec 2020	Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2020, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.	Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table The payer's own page
Off-the-shelf speech recognition training datasets (wake words, read sentences, dialogues; Chinese and foreign languages), usage licences sold by Haitian Ruisheng in 2019: 136,411 hours	549 元/小时 (CNY 549 an hour, average over 136,411 hours sold) per hour, average across all speech recognition dataset hours licensed in 2019	Dec 2019	Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2019, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate. The same filing splits 2019 by language: Chinese 55,957 hours at CNY 317 an hour, foreign languages 80,728 hours at CNY 793 an hour.	Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table The payer's own page
Off-the-shelf speech synthesis (TTS) training datasets, usage licences sold by Haitian Ruisheng in 2019: 274 hours	24,946 元/小时 (CNY 24,946 an hour, average over 274 hours sold) per hour, average across all speech synthesis dataset hours licensed in 2019	Dec 2019	Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2019, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.	Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table The payer's own page
Off-the-shelf speech recognition training datasets (wake words, read sentences, dialogues; Chinese and foreign languages), usage licences sold by Haitian Ruisheng in 2018: 115,636 hours	504 元/小时 (CNY 504 an hour, average over 115,636 hours sold) per hour, average across all speech recognition dataset hours licensed in 2018	Dec 2018	Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2018, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.	Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table The payer's own page
Off-the-shelf speech synthesis (TTS) training datasets, usage licences sold by Haitian Ruisheng in 2018: 96 hours	20,124 元/小时 (CNY 20,124 an hour, average over 96 hours sold) per hour, average across all speech synthesis dataset hours licensed in 2018	Dec 2018	Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2018, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.	Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table The payer's own page

Custom speech data collection to a buyer's spec (recruited speakers, recorded, transcribed and QA'd), 4 to 12 weeks
Price
$50 to $300 per hour of delivered audio
per hour of delivered audio
When
Oct 2026
What to know
AIxBlock's range after transcription and QA, set by language scarcity, accent diversity and annotation depth; the same range appears in its call-center guide (held under aixblock-custom-1000-hours). The page also warns that quotes below $20 an hour for custom collection usually signal unsustainable contributor rates. The page carries no date (footer 2025).
Source
AIxBlock, How to Choose a Speech Data Collection Service for ASR
A vendor's published price guide
Studio-quality text-to-speech (TTS) voice data recorded to order
Price
$200 to $500 per hour
per hour of audio
When
Oct 2026
What to know
AIxBlock puts studio-quality TTS data above its $50 to $300 range for ordinary custom speech collection. The page carries no date (footer 2025).
Source
AIxBlock, How to Choose a Speech Data Collection Service for ASR
A vendor's published price guide
Custom on-site speech and audio recordings made to the buyer's requirements by field teams (Spire Light on Datarade)
Price
Usage-based $250 / Audio hour
per hour of audio recorded to order
When
Oct 2026
What to know
A 'starts at' price on a data marketplace; Spire Light is the company behind SpeechData.ai (its footer reads 'A Spirelight company'), and detailed pricing is by request. Delivered as .json and .wav with human QA.
Source
Datarade, Custom On-Site Speech & Audio Data Collection (Spire Light)
A catalogue or marketplace listing
Custom paired conversational speech recordings (audio, optionally video) built to project requirements in 80+ languages (Spire Light on Datarade)
Price
Usage-based $90 / Audio hour
per hour of audio recorded to order
When
Oct 2026
What to know
A 'starts at' price on a data marketplace for paired conversations recorded to order from a network of over 30,000 people; the $250 on-site listing from the same seller is for field recordings. Detailed pricing is by request.
Source
Datarade, Custom Pair Conversational Speech Audio / Video Data Collection Dataset in 80+ Languages (Spire Light)
A catalogue or marketplace listing
Median advertised hourly pay for voice acting and voice-over roles in AI training
Price
$35 an hour (live median)
per hour, the median of 111 advertised rates
When
Oct 2026
What to know
The median of 111 distinct advertised hourly rates across 8 platforms; the typical range is $35 to $75 and the top is $150 at Mercor. AITraining.jobs has the same owner as AIData.Broker.
Source
AITraining.jobs, Voice Acting and Voice Over Jobs for AI
A catalogue or marketplace listing
Nexdata Indonesian spontaneous dialogue smartphone speech, 747 hours from 1,190 native speakers, transcribed with PII redacted (on Datarade)
Price
Starts at $20,000 / purchase (one-off)
per dataset of 747 hours
When
Oct 2026
What to know
A 'starts at' one-off licence price, the same $20,000 floor Nexdata shows across its Datarade speech products; spontaneous two-party dialogue at this floor works out cheaper per hour than Nexdata's scripted Indonesian set.
Source
Datarade, Nexdata Indonesian Spontaneous Dialogue Smartphone Speech Dataset, 747 Hours
A catalogue or marketplace listing
Nexdata Indonesian scripted monologue smartphone speech, 359 hours from 496 speakers, transcribed (on Datarade)
Price
Starts at $20,000 / purchase (one-off)
per dataset of 359 hours
When
Oct 2026
What to know
A 'starts at' one-off licence price; Nexdata shows the same $20,000 floor on many of its Datarade speech products, so it reads as a minimum order rather than a priced quote. The same corpus lists on ELRA at EUR 57,978.50 for commercial use, about 2.6 times this floor.
Source
Datarade, Nexdata Indonesian Speech Data by Mobile Phone, 359 Hours
A catalogue or marketplace listing
REMI Data licensed native-speaker speech recordings in 13+ languages including Yoruba, Amharic and Swahili, WAV 16 kHz with verbatim transcripts and perpetual commercial consent (on Datarade)
Price
Usage-based $120 / per hour (one-off purchase from $500)
per hour of audio
When
Oct 2026
What to know
One per-hour rate across all the languages the listing names, Yoruba among them; the stated volume is 100 hours for the whole product, so how many Yoruba hours exist off the shelf is not stated, and custom collection is offered beside it.
Source
Datarade, REMI Data Multilingual Speech Recordings, 13+ Languages
A catalogue or marketplace listing
IARPA Babel Vietnamese Language Pack IARPA-babel107b-v0.7 (LDC2017S01): about 201 hours of Vietnamese conversational and scripted telephone speech with transcripts, collected by Appen in 2012
Price
$25.00 (non-member fee)
per corpus of about 201 hours
When
Oct 2026
What to know
LDC's non-member fee, shown under View Fees (the reduced-licence fee is also $25.00). The Babel packs were paid for by the US government and are priced near zero; non-member LDC licences exclude commercial use, so this is a research price, not a market price for Vietnamese speech.
Source
LDC catalog, IARPA Babel Vietnamese Language Pack IARPA-babel107b-v0.7 (LDC2017S01)
A catalogue or marketplace listing
NaijaVoices suite of speech-text datasets (Hausa, Igbo and Yoruba; 1,800+ hours from 5,000+ speakers): the Supporter tier that adds a waiver for commercial use
Price
Donate $5,000 or More
per commercial-use waiver for the suite of 1,800+ hours
When
Oct 2026
What to know
A community project's minimum donation for commercial rights, not a negotiated price; the default licence (CC BY-NC-SA 4.0) is free for non-commercial use. Read speech of expert-written sentences, so it sits far below conversational listings.
Source
NaijaVoices, Membership tiers
A catalogue or marketplace listing
NaijaVoices: total cost of creating its 1,800+ hour speech-text dataset in Hausa, Igbo and Yoruba
Price
substantial (>$600k)
per dataset of 1,800+ hours
When
Oct 2026
What to know
The producer's own statement of what the dataset cost to make (sentence writing, recording, double review), set against its $5,000 commercial waiver: the commercial right sells at about 1 percent of production cost.
Source
NaijaVoices, Membership FAQ
The payer's own page
NaijaVoices: the rate it costs per hour of recording by its Nigerian 'data farmers' (Hausa, Igbo, Yoruba read speech)
Price
rate of $30/hour for recording
per hour of recording
When
Oct 2026
What to know
The rate NaijaVoices uses to turn a donation into hours of speech data; the page also states $0.12 per sentence written, a double review of all output, each sentence recorded twice and at most 5 hours per speaker.
Source
NaijaVoices, Membership: donation impact calculation breakdown
The payer's own page
SpeechData.ai Indonesian conversational speech dataset: 2,000 hours of spontaneous two-person conversations recorded in Indonesia, 200 native speakers, dual-channel, time-aligned transcripts
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf asking price for a commercial training and evaluation licence with speaker consent documentation; the page says volume discounts are available for full-dataset and multi-language purchases.
Source
SpeechData.ai, Indonesian conversational speech dataset
A catalogue or marketplace listing
SpeechData.ai Vietnamese conversational speech dataset: 2,000 hours of spontaneous two-person conversations recorded in Vietnam, 200 native speakers, dual-channel, time-aligned transcripts
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf asking price for a commercial training and evaluation licence with speaker consent documentation; the page says volume discounts are available for full-dataset and multi-language purchases.
Source
SpeechData.ai, Vietnamese conversational speech dataset
A catalogue or marketplace listing
Off-the-shelf speech data, major languages
Price
$50-100 per hour of audio
per hour
When
Oct 2026
What to know
A vendor's price table; low-resource languages are $60 to $150 on the same table, custom collection $150 to $400+.
Source
SpeechData.ai, How to buy speech data (price table)
A vendor's published price guide
Recorded speech sold off the shelf
Price
$60-150 per hour of audio
per hour
When
Oct 2026
What to know
A vendor's account of the market for off-the-shelf conversational speech; it raised the floor from $25 (read 11 October 2026) to $60. Its own catalogue sits at $60 to $95.
Source
SpeechData.ai, How to buy speech data (FAQ)
A vendor's published price guide
AI Tutor specialised in multilingual audio, Yoruba (xAI): records, annotates and transcribes Yoruba speech and localises English text into Yoruba
Price
$35/hour - $45/hour
per hour (paid to the tutor, US-based candidates)
When
Oct 2026
What to know
The rate is stated for US-based candidates only; the post says international candidates get their rate during recruitment, so a Yoruba speaker in Nigeria is likely paid differently (not stated). Same US band as xAI's other AI Tutor posts (French, Hausa). Posted 2026-08-07, updated 2026-10-13; contractor hours vary, and projects may require at least 10 hours a week.
Source
xAI job board (Greenhouse), AI Tutor - Yoruba
The payer's own page
Custom 1,000-hour English call-center speech collection with accent coverage and basic transcription, 12 weeks
Price
$80,000 to $200,000
per 1,000-hour project
When
Oct 2026
What to know
Same guide prices custom collection at $50 to $300 per delivered hour, versus $5,000 to $25,000 for the same volume off the shelf.
Source
AIxBlock
A vendor's published price guide
Off-the-shelf call-center audio licensed from existing corpora (AIxBlock guide)
Price
$5 to $25 per hour of audio
per hour of audio
When
Oct 2026
What to know
Undated vendor guide read on this date; English call-center audio sits at the low end, low-resource languages and regulated domains at the high end.
Source
AIxBlock
A vendor's published price guide
Sound-effects catalogue for AI training: 350+ libraries, 53,000 files (Krampfstadt Studio), one-off purchase
Price
€60,000 / purchase
per catalogue of 53k sound files
When
Oct 2026
What to know
The same seller asks $78,900 for this catalogue on Opendatabay; a yearly licence is EUR 20,000 and a monthly one EUR 3,000.
Source
Datarade, Krampfstadt Studio 350+ SFX libraries, 53k audio files
A catalogue or marketplace listing
Sound-effects catalogue for AI training: 350+ libraries, 53,000 files (Krampfstadt Studio), yearly licence
Price
€20,000 / year
per year
When
Oct 2026
What to know
Three years of the yearly licence cost about what the one-off purchase does (EUR 60,000).
Source
Datarade, Krampfstadt Studio 350+ SFX libraries, 53k audio files
A catalogue or marketplace listing
Two-speaker spontaneous conversations collected on demand in rare and low-resource languages and accents, with transcripts (Silencio Network)
Price
Usage-based $50 / Hour
per hour of audio collected to order
When
Oct 2026
What to know
Collection to order from a crowd of contributors in 180+ countries, priced per delivered hour, with discounts above 1,000 hours.
Source
Datarade, Silencio Network Spontaneous Conversations On-Demand (Accent and Dialect Focus)
A catalogue or marketplace listing
Real customer-service calls in English and Hindi for chatbot and voicebot training (WiserBrand), usage-based licence
Price
Usage-based $25 / hour
per hour of audio (usage-based licence)
When
Oct 2026
What to know
Same per-hour price as the seller's US English calls: no premium for Hindi or code-switched speech at this seller.
Source
Datarade, WiserBrand Chatbot Conversation Dataset (English and Hindi)
A catalogue or marketplace listing
Real US customer-service call recordings, speaker-separated 48 kHz, PII-redacted (WiserBrand, 40,000+ hours available), usage-based licence
Price
Usage-based $25 / hour (one-off purchase from $10,000)
per hour of audio (usage-based licence)
When
Oct 2026
What to know
Seller's own per-hour list price on the Datarade marketplace for real (not staged) support calls; a one-off purchase starts at $10,000.
Source
Datarade, WiserBrand Real Call Center Conversations for ASR and Voice AI
A catalogue or marketplace listing
ALLIES corpus: 900 hours of French broadcast news for diarization evaluation (ELRA catalogue ELRA-S0486)
Price
31250.00 € (non-member, commercial use)
per corpus of 900 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Broadcast speech in volume, so cheap per hour. A non-commercial-use licence is EUR 2,500 for an academic and EUR 9,375 for a company.
Source
ELRA catalogue, ALLIES Corpus (ELRA-S0486)
A catalogue or marketplace listing
American children's read speech (storybooks, songs), microphone, 50.5 hours from 219 children (ELRA catalogue ELRA-S0468)
Price
28785.00 € (non-member, commercial use)
per corpus of 50.5 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech: about three and a half times the per-hour price of the adult American English set in the same series (ELRA-S0467). ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, American Children Speech Data by Microphone - 50 Hours (ELRA-S0468)
A catalogue or marketplace listing
American English prompted speech recorded on mobile phones, 800 hours from 1,842 speakers (ELRA catalogue ELRA-S0437)
Price
136800.00 € (non-member, commercial use)
per corpus of 800 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, American English Speech Data by Mobile Phone - 800 Hours (ELRA-S0437)
A catalogue or marketplace listing
American English read speech recorded on mobile phones, 215 hours from 349 speakers (ELRA catalogue ELRA-S0467)
Price
34722.50 € (non-member, commercial use)
per corpus of 215 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, American English Speech Data by Mobile Phone_Reading - 215 Hours (ELRA-S0467)
A catalogue or marketplace listing
Italian Speech Corpus 1 (Appen): 44 hours from 202 speakers, 4 microphone channels, office and public places (ELRA catalogue ELRA-S0147)
Price
15000.00 € (non-member, commercial use)
per corpus of 44 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. An Appen off-the-shelf corpus distributed through ELRA, one of the few Appen corpora with a public price. A non-commercial-use licence is EUR 1,500 for an academic and EUR 15,000 for a company.
Source
ELRA catalogue, Italian Speech Corpus 1 (Appen) (ELRA-S0147)
A catalogue or marketplace listing
Spanish Speech Corpus 1 (Appen): 39 hours from 200 speakers, 4 microphone channels, office and public places (ELRA catalogue ELRA-S0149)
Price
15000.00 € (non-member, commercial use)
per corpus of 39 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. An Appen off-the-shelf corpus distributed through ELRA. A non-commercial-use licence is EUR 1,500 for an academic and EUR 15,000 for a company.
Source
ELRA catalogue, Spanish Speech Corpus 1 (Appen) (ELRA-S0149)
A catalogue or marketplace listing
Brazilian Portuguese prompted speech recorded on mobile phones, 1,044 hours from 2,038 speakers (ELRA catalogue ELRA-S0445)
Price
247950.00 € (non-member, commercial use)
per corpus of 1,044 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Brazilian Portuguese Speech Data by Mobile Phone - 1,044 Hours (ELRA-S0445)
A catalogue or marketplace listing
BREF-120: 100 hours of French read speech (ELRA catalogue ELRA-S0067)
Price
15000.00 € (non-member, commercial use)
per corpus of 100 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. A classic read-speech corpus. A non-commercial-use licence is EUR 4,000 for an academic and EUR 15,000 for a company.
Source
ELRA catalogue, BREF-120 - A large corpus of French read speech (ELRA-S0067)
A catalogue or marketplace listing
British children's read speech (textbooks, storybooks), high-fidelity microphone, 55 hours from 201 children (ELRA catalogue ELRA-S0474)
Price
39187.50 € (non-member, commercial use)
per corpus of 55 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech: about three times the per-hour price of the adult British English set in the same series (ELRA-S0466). ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, British Children Speech Data by Microphone - 55 Hours (ELRA-S0474)
A catalogue or marketplace listing
British English prompted speech recorded on mobile phones, 831 hours from 1,651 speakers (ELRA catalogue ELRA-S0448)
Price
213151.50 € (non-member, commercial use)
per corpus of 831 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, British English Speech Data by Mobile Phone - 831 Hours (ELRA-S0448)
A catalogue or marketplace listing
British English read speech recorded on mobile phones, 199 hours from 346 speakers (ELRA catalogue ELRA-S0466)
Price
47262.50 € (non-member, commercial use)
per corpus of 199 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, British English Speech Data by Mobile Phone_Reading - 199 Hours (ELRA-S0466)
A catalogue or marketplace listing
Cantonese Speecon: about 231 hours of adult and child speech over 4 microphone channels in office, car and public places (ELRA catalogue ELRA-S0287)
Price
75000.00 € (non-member, commercial use)
per corpus of 231 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Multi-microphone, multi-environment recordings. A non-commercial-use licence is EUR 60,000 for an academic and EUR 75,000 for a company.
Source
ELRA catalogue, Cantonese Speecon database (ELRA-S0287)
A catalogue or marketplace listing
Chinese Kids Speech database (Lower Grade): 87 hours from 184 children aged 6 to 10, smartphone (ELRA catalogue ELRA-S0496)
Price
15400.00 € (non-member, commercial use)
per corpus of 87 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech. A non-commercial-use licence is EUR 2,600 for an academic and EUR 10,300 for a company.
Source
ELRA catalogue, Chinese Kids Speech database (Lower Grade) (ELRA-S0496)
A catalogue or marketplace listing
Chinese Kids Speech database (Upper Grade): 72 hours from 161 children aged 10 to 12, smartphone (ELRA catalogue ELRA-S0497)
Price
14000.00 € (non-member, commercial use)
per corpus of 72 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech. A non-commercial-use licence is EUR 2,300 for an academic and EUR 9,300 for a company.
Source
ELRA catalogue, Chinese Kids Speech database (Upper Grade) (ELRA-S0497)
A catalogue or marketplace listing
CLE Pakistan Urdu Speech Corpus: 61.62 hours of phonetically rich read Urdu on varied microphones (ELRA catalogue ELRA-S0403)
Price
23400.00 € (non-member, commercial use)
per corpus of 61.62 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Read Urdu here costs several times conversational Urdu in commercial catalogues. A non-commercial-use licence is EUR 15,600 for an academic and EUR 23,400 for a company.
Source
ELRA catalogue, CLE Pakistan Urdu Speech Corpus (ELRA-S0403)
A catalogue or marketplace listing
Corpus of Spontaneous Japanese (CSJ), 650 hours as listed by ELRA, distributed for NINJAL (ELRA catalogue ELRA-S0488)
Price
43125.00 € (non-member, commercial use)
per corpus of 650 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. The owner, NINJAL, licenses the same corpus directly for 5 million yen (open-ended, up to 10 users), well under ELRA's price. A non-commercial-use licence is EUR 431.25 for an academic and EUR 43,125 for a company.
Source
ELRA catalogue, Corpus of Spontaneous Japanese (CSJ) (ELRA-S0488)
A catalogue or marketplace listing
Danish SpeechDat(II) FDB-4000: 242 hours of fixed-network telephone speech from 4,000 speakers (ELRA catalogue ELRA-S0073)
Price
56000.00 € (non-member, commercial use)
per corpus of 242 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Fixed-network telephone speech from 4,000 speakers. A non-commercial-use licence is EUR 48,000 for an academic and EUR 56,000 for a company.
Source
ELRA catalogue, Danish SpeechDat(II) FDB-4000 (ELRA-S0073)
A catalogue or marketplace listing
Dutch Polyphone Database: 335 hours of Dutch and Flemish telephone speech (ELRA catalogue ELRA-S0010)
Price
35000.00 € (non-member, commercial use)
per corpus of 335 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Telephone speech. A non-commercial-use licence is EUR 20,000 for an academic and EUR 35,000 for a company.
Source
ELRA catalogue, Dutch Polyphone Database (ELRA-S0010)
A catalogue or marketplace listing
Egyptian Arabic Speecon: 196 hours of Modern Standard Arabic as spoken in Egypt, adults and children, 4 microphone channels in office, car and public places (ELRA catalogue ELRA-S0308)
Price
75000.00 € (non-member, commercial use)
per corpus of 196 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Multi-microphone, multi-environment recordings. A non-commercial-use licence is EUR 60,000 for an academic and EUR 75,000 for a company.
Source
ELRA catalogue, Egyptian Arabic Speecon database (ELRA-S0308)
A catalogue or marketplace listing
ESTER 2: about 100 hours of transcribed French broadcast news (ELRA catalogue ELRA-S0338)
Price
25000.00 € (non-member, commercial use)
per corpus of 100 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Commercial use costs several times the non-commercial price. A non-commercial-use licence is EUR 2,000 for an academic and EUR 7,500 for a company.
Source
ELRA catalogue, ESTER 2 Corpus (ELRA-S0338)
A catalogue or marketplace listing
EthioSpeech: 391 hours of read speech in six Ethiopian languages (Amharic, Tigrigna, Oromo, Somali, Afar, Sidama) (ELRA catalogue ELRA-S0494)
Price
5400.00 € (non-member, commercial use)
per corpus of 391 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Low-resource African languages at a low commercial price per hour. A non-commercial-use licence is free for an academic and EUR 5,400 for a company.
Source
ELRA catalogue, EthioSpeech (ELRA-S0494)
A catalogue or marketplace listing
French prompted speech recorded on mobile phones, 769 hours from 1,623 speakers (ELRA catalogue ELRA-S0446)
Price
182637.50 € (non-member, commercial use)
per corpus of 769 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, French Speech Data by Mobile Phone - 769 Hours (ELRA-S0446)
A catalogue or marketplace listing
French read speech recorded on mobile phones (France, Canada, Africa), 231.9 hours from 406 speakers (ELRA catalogue ELRA-S0465)
Price
50670.15 € (non-member, commercial use)
per corpus of 231.9 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, French Speech Data by Mobile Phone_Reading - 231 Hours (ELRA-S0465)
A catalogue or marketplace listing
English spoken by native German speakers (accented English), mobile phones, 535 hours from 1,162 speakers (ELRA catalogue ELRA-S0442)
Price
152475.00 € (non-member, commercial use)
per corpus of 535 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Accented English priced at the top tier, the same as native German speech in the series. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, German Speaking English Speech Data by Mobile Phone - 535 Hours (ELRA-S0442)
A catalogue or marketplace listing
German prompted speech recorded on mobile phones (generic, in-car, smart-home scripts), 1,796.70 hours from 3,442 speakers (ELRA catalogue ELRA-S0449)
Price
512059.50 € (non-member, commercial use)
per corpus of 1,796.70 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Top tier of a per-hour price ladder in this series (EUR 285 an hour). ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, German Speech Data by Mobile Phone - 1,796 Hours (ELRA-S0449)
A catalogue or marketplace listing
German read speech recorded on mobile phones, 211 hours from 327 speakers (ELRA catalogue ELRA-S0473)
Price
46103.50 € (non-member, commercial use)
per corpus of 211 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, German Speech Data by Mobile Phone_Reading - 211 Hours (ELRA-S0473)
A catalogue or marketplace listing
GlobalPhone Hausa: 8.73 hours of read Hausa speech (ELRA catalogue ELRA-S0347)
Price
3600.00 € (non-member, commercial use)
per corpus of 8.73 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. A small research corpus; per hour it is far dearer than larger commercial Hausa sets. A non-commercial-use licence is EUR 700 for an academic and EUR 3,600 for a company.
Source
ELRA catalogue, GlobalPhone Hausa (ELRA-S0347)
A catalogue or marketplace listing
GlobalPhone Tamil: 20 hours of read Tamil speech (ELRA catalogue ELRA-S0205)
Price
600.00 € (non-member, commercial use)
per corpus of 20 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Other GlobalPhone languages list at EUR 3,600 for about 20 hours, six times this price. A non-commercial-use licence is EUR 125 for an academic and EUR 600 for a company.
Source
ELRA catalogue, GlobalPhone Tamil (ELRA-S0205)
A catalogue or marketplace listing
Gram Vaani: 130 hours of Hindi phone recordings from 4,000 mostly rural speakers in Bihar, Jharkhand and Madhya Pradesh (ELRA catalogue ELRA-S0405)
Price
50000.00 € (non-member, commercial use)
per corpus of 130 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Rural, accented, real-world telephone speech from a community voice platform. A non-commercial-use licence is free for an academic and EUR 50,000 for a company.
Source
ELRA catalogue, Gram Vaani data set (ELRA-S0405)
A catalogue or marketplace listing
Greek SpeechDat(II) FDB-5000: 422 hours of fixed-network telephone speech from 5,000 speakers (ELRA catalogue ELRA-S0118)
Price
70000.00 € (non-member, commercial use)
per corpus of 422 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Telephone speech. A non-commercial-use licence is EUR 60,000 for an academic and EUR 70,000 for a company.
Source
ELRA catalogue, Greek SpeechDat(II) FDB-5000 (ELRA-S0118)
A catalogue or marketplace listing
Hindi prompted speech recorded on mobile phones, 759 hours from 1,425 speakers (ELRA catalogue ELRA-S0452)
Price
115368.00 € (non-member, commercial use)
per corpus of 759 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Bottom tier of the series' per-hour price ladder (EUR 152 an hour), shared with Indian English. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Hindi Speech Data by Mobile Phone - 759 Hours (ELRA-S0452)
A catalogue or marketplace listing
Hindi read speech recorded on mobile phones, 240 hours from 401 speakers, transcribed (ELRA catalogue ELRA-S0463)
Price
34200.00 € (non-member, commercial use)
per corpus of 240 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Hindi Speech Data by Mobile Phone_R - 240 Hours (ELRA-S0463)
A catalogue or marketplace listing
Indonesian read speech recorded on mobile phones, 359 hours from 496 speakers (ELRA catalogue ELRA-S0470)
Price
57978.50 € (non-member, commercial use)
per corpus of 359 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Indonesian Speech Data by Mobile Phone_R - 359 Hours (ELRA-S0470)
A catalogue or marketplace listing
Japanese Kids Speech database (Lower Grade): 116.60 hours from 179 children aged 6 to 9, smartphone (ELRA catalogue ELRA-S0411)
Price
12300.00 € (non-member, commercial use)
per corpus of 116.60 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Children's speech, here priced below the adult Japanese sets. A non-commercial-use licence is EUR 2,100 for an academic and EUR 8,200 for a company.
Source
ELRA catalogue, Japanese Kids Speech database (Lower Grade) (ELRA-S0411)
A catalogue or marketplace listing
Japanese read speech recorded on mobile phones, 234.8 hours from 799 speakers (ELRA catalogue ELRA-S0476)
Price
35689.60 € (non-member, commercial use)
per corpus of 234.8 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Japanese Speech Data by Mobile Phone_R - 234 Hours (ELRA-S0476)
A catalogue or marketplace listing
Korean read speech recorded on mobile phones, 197 hours from 291 speakers (ELRA catalogue ELRA-S0475)
Price
31815.50 € (non-member, commercial use)
per corpus of 197 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Korean Speech Data by Mobile Phone_Reading - 197 Hours (ELRA-S0475)
A catalogue or marketplace listing
LILA Korean: 81.16 hours of mobile-network telephone speech from 1,000 speakers (ELRA catalogue ELRA-S0295)
Price
51875.00 € (non-member, commercial use)
per corpus of 81.16 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Mobile telephone speech in chosen acoustic conditions. A non-commercial-use licence is EUR 46,405 for an academic and EUR 51,875 for a company.
Source
ELRA catalogue, LILA Korean database (ELRA-S0295)
A catalogue or marketplace listing
MEDIA speech database: 70 hours of French human-machine dialogue (hotel booking scenario) (ELRA catalogue ELRA-S0272)
Price
10000.00 € (non-member, commercial use)
per corpus of 70 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Produced under a French government research programme. A non-commercial-use licence is free for an academic and EUR 10,000 for a company.
Source
ELRA catalogue, MEDIA speech database for French (ELRA-S0272)
A catalogue or marketplace listing
MoveOn Speech and Noise Corpus: 6 hours of British English command speech recorded on motorcycles (ELRA catalogue ELRA-S0370)
Price
9000.00 € (non-member, commercial use)
per corpus of 6 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Noisy speech for a niche domain (command speech on motorcycles, with a focus on police use); the dearest per hour of the ELRA prices recorded here. A non-commercial-use licence is EUR 1,500 for an academic and EUR 9,000 for a company.
Source
ELRA catalogue, MoveOn Speech and Noise Corpus (ELRA-S0370)
A catalogue or marketplace listing
OrienTel Egypt: colloquial Egyptian Arabic over fixed and mobile telephone networks, 50 hours from 750 speakers (ELRA catalogue ELRA-S0221)
Price
60250.00 € (non-member, commercial use)
per corpus of 50 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Egyptian Arabic here costs about 20 times SpeechData.ai's $65 an hour for Egyptian conversation. A non-commercial-use licence is EUR 45,187 for an academic and EUR 60,250 for a company.
Source
ELRA catalogue, OrienTel Egypt MCA (Modern Colloquial Arabic) database (ELRA-S0221)
A catalogue or marketplace listing
OrienTel Jordan: colloquial Arabic over fixed and mobile telephone networks, 50 hours from 757 speakers (ELRA catalogue ELRA-S0289)
Price
60250.00 € (non-member, commercial use)
per corpus of 50 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Dialect Arabic telephone speech at over EUR 1,200 an hour. A non-commercial-use licence is EUR 45,187 for an academic and EUR 60,250 for a company.
Source
ELRA catalogue, OrienTel Jordan MCA (Modern Colloquial Arabic) database (ELRA-S0289)
A catalogue or marketplace listing
OrienTel Tunisia: colloquial Arabic over fixed and mobile telephone networks, 50 hours from 792 speakers (ELRA catalogue ELRA-S0186)
Price
27500.00 € (non-member, commercial use)
per corpus of 50 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Dialect Arabic telephone speech. A non-commercial-use licence is EUR 16,500 for an academic and EUR 27,500 for a company.
Source
ELRA catalogue, OrienTel Tunisia MCA (Modern Colloquial Arabic) database (ELRA-S0186)
A catalogue or marketplace listing
Russian prompted speech recorded on mobile phones, 1,002 hours from 1,960 speakers (ELRA catalogue ELRA-S0443)
Price
180861.00 € (non-member, commercial use)
per corpus of 1,002 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Russian Speech Data by Mobile Phone - 1,002 Hours (ELRA-S0443)
A catalogue or marketplace listing
Spanish emotional speech synthesis database: 4 hours from two professional speakers in neutral and six emotional styles (ELRA catalogue ELRA-S0329)
Price
4000.00 € (non-member, commercial use)
per corpus of 4 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Emotional, acted speech for text-to-speech. A non-commercial-use licence is free for an academic and EUR 4,000 for a company.
Source
ELRA catalogue, Emotional speech synthesis database (ELRA-S0329)
A catalogue or marketplace listing
Spanish read speech recorded on mobile phones (Spain, Mexico, Venezuela), 227 hours, transcribed (ELRA catalogue ELRA-S0464)
Price
40973.50 € (non-member, commercial use)
per corpus of 227 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Spanish Speech Data by Mobile Phone_R - 227 Hours (ELRA-S0464)
A catalogue or marketplace listing
Castilian Spanish SpeechDat(II) FDB-4000: 246.67 hours of fixed-network telephone speech from 4,000 speakers (ELRA catalogue ELRA-S0102)
Price
19500.00 € (non-member, commercial use)
per corpus of 246.67 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Telephone speech. A non-commercial-use licence is EUR 12,500 for an academic and EUR 19,500 for a company.
Source
ELRA catalogue, Spanish SpeechDat(II) FDB-4000 (ELRA-S0102)
A catalogue or marketplace listing
Swiss-German SpeechDat(II) FDB-2000: 110 hours of fixed-network telephone speech from 2,000 speakers (ELRA catalogue ELRA-S0105)
Price
38000.00 € (non-member, commercial use)
per corpus of 110 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. A dialect with few corpora. A non-commercial-use licence is EUR 26,000 for an academic and EUR 38,000 for a company.
Source
ELRA catalogue, Swiss-German SpeechDat(II) FDB-2000 (ELRA-S0105)
A catalogue or marketplace listing
TC-STAR English training corpus: about 290 hours of European Parliament plenary speeches (ELRA catalogue ELRA-S0251)
Price
9000.00 € (non-member, commercial use)
per corpus of 290 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Public parliamentary speech, priced well below collected speech. A non-commercial-use licence is EUR 6,270 for an academic and EUR 9,000 for a company.
Source
ELRA catalogue, TC-STAR English Training Corpora for ASR: Recordings of EPPS Speech (ELRA-S0251)
A catalogue or marketplace listing
TC-STAR Spanish baseline female voice: 10 hours of one speaker reading in a noise-reduced room, for speech synthesis (ELRA catalogue ELRA-S0309)
Price
7800.00 € (non-member, commercial use)
per corpus of 10 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Single-speaker studio voice data for text-to-speech. A non-commercial-use licence is EUR 3,850 for an academic and EUR 7,800 for a company.
Source
ELRA catalogue, TC-STAR Spanish Baseline Female Speech Database (ELRA-S0309)
A catalogue or marketplace listing
Thai read speech recorded on mobile phones, 203 hours from 498 speakers (ELRA catalogue ELRA-S0471)
Price
47093.40 € (non-member, commercial use)
per corpus of 203 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Subsets can be requested. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Thai Speech Data by Mobile Phone_Reading - 203 Hours (ELRA-S0471)
A catalogue or marketplace listing
THERADIA WoZ corpus: 86 hours of audiovisual sessions for detecting affect in home cognitive-training (healthcare) (ELRA catalogue ELRA-S0499)
Price
6000.00 € (non-member, commercial use)
per corpus of 86 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Healthcare-domain affective speech. A non-commercial-use licence is free for an academic and EUR 6,000 for a company.
Source
ELRA catalogue, THERADIA WoZ Corpus (ELRA-S0499)
A catalogue or marketplace listing
TRAD Pashto Broadcast News: 108 hours of transcribed TV and radio news from more than 1,000 speakers (ELRA catalogue ELRA-S0381)
Price
28000.00 € (non-member, commercial use)
per corpus of 108 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. A low-resource language, priced well above broadcast speech in big languages. A non-commercial-use licence is EUR 3,500 for an academic and EUR 28,000 for a company.
Source
ELRA catalogue, TRAD Pashto Broadcast News Speech Corpus (ELRA-S0381)
A catalogue or marketplace listing
UK English Speecon: 204 hours from 606 adults and 51 children, 4 microphone channels in office, car and public places (ELRA catalogue ELRA-S0215)
Price
75000.00 € (non-member, commercial use)
per corpus of 204 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Every Speecon corpus seen in this catalogue lists at EUR 75,000, whatever the language. A non-commercial-use licence is EUR 60,000 for an academic and EUR 75,000 for a company.
Source
ELRA catalogue, UK English Speecon database (ELRA-S0215)
A catalogue or marketplace listing
Vietnamese prompted speech recorded on mobile phones, 760 hours from 1,751 speakers (ELRA catalogue ELRA-S0432)
Price
122740.00 € (non-member, commercial use)
per corpus of 760 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. Part of a series priced by language at a fixed rate per hour. ELRA lists no non-commercial licence for it.
Source
ELRA catalogue, Vietnamese Speech Data by Mobile Phone - 760 Hours (ELRA-S0432)
A catalogue or marketplace listing
Welsh SpeechDat(II) FDB-2000: 170 hours of Welsh telephone speech (ELRA catalogue ELRA-S0075)
Price
18000.00 € (non-member, commercial use)
per corpus of 170 hours
When
Oct 2026
What to know
ELRA's non-member price for a company under the commercial-use (VAR) licence. A minority European language. A non-commercial-use licence is EUR 11,000 for an academic and EUR 18,000 for a company.
Source
ELRA catalogue, Welsh SpeechDat(II) FDB-2000 (ELRA-S0075)
A catalogue or marketplace listing
AIDA Ukrainian Broadcast and Telephone Speech: about 156 hours with 1.2 million words of transcripts
Price
$2,000.00 (non-member fee)
per corpus of about 156 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Source
LDC catalog, AIDA Ukrainian Broadcast and Telephone Speech Audio and Transcripts (LDC2023S01)
A catalogue or marketplace listing
AISHELL-1: about 520 hours of Mandarin read speech from 400 speakers on three devices
Price
$3,000.00 (non-member fee)
per corpus of about 520 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, AISHELL-1 (LDC2018S14)
A catalogue or marketplace listing
AnnoDIFP Session Audio and Transcripts: 438.34 hours of English in-person interview audio with personality scores
Price
$3,000.00 (non-member fee)
per corpus of about 438.34 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, AnnoDIFP Session Audio and Transcripts (LDC2025S06)
A catalogue or marketplace listing
Air Traffic Control Complete: about 70 hours of controller and pilot radio transmissions
Price
$1,150.00 (non-member fee)
per corpus of about 70 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $575.00.
Source
LDC catalog, Air Traffic Control Complete (LDC94S14A)
A catalogue or marketplace listing
BOLT CTS CALLFRIEND CALLHOME Egyptian Arabic Audio: about 116 hours of telephone conversations
Price
$2,000.00 (non-member fee)
per corpus of about 116 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Source
LDC catalog, BOLT CTS CALLFRIEND CALLHOME Egyptian Arabic Audio (LDC2025S09)
A catalogue or marketplace listing
Call My Net 1: 364 hours of conversational telephone speech in Tagalog, Cebuano, Cantonese and Mandarin
Price
$2,500.00 (non-member fee)
per corpus of about 364 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.
Source
LDC catalog, Call My Net 1 (LDC2024S05)
A catalogue or marketplace listing
CALLHOME American English Second Edition: about 56 hours of unscripted telephone conversations
Price
$3,000.00 (non-member fee)
per corpus of about 56 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.
Source
LDC catalog, CALLHOME American English Second Edition (LDC2026S08)
A catalogue or marketplace listing
CALLHOME German Second Edition: about 48 hours of unscripted telephone conversations
Price
$3,000.00 (non-member fee)
per corpus of about 48 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.
Source
LDC catalog, CALLHOME German Second Edition (LDC2026S06)
A catalogue or marketplace listing
CALLHOME Japanese Second Edition: about 49 hours of unscripted telephone conversations
Price
$3,000.00 (non-member fee)
per corpus of about 49 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, CALLHOME Japanese Second Edition (LDC2026S02)
A catalogue or marketplace listing
CALLHOME Mandarin Chinese Second Edition: about 38 hours of unscripted telephone conversations
Price
$3,000.00 (non-member fee)
per corpus of about 38 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.
Source
LDC catalog, CALLHOME Mandarin Chinese Second Edition (LDC2026S11)
A catalogue or marketplace listing
CALLHOME Spanish Second Edition: about 38 hours of unscripted telephone conversations
Price
$3,000.00 (non-member fee)
per corpus of about 38 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, CALLHOME Spanish Second Edition (LDC2026S04)
A catalogue or marketplace listing
Fisher English Training Speech Part 1: 984 hours of English conversational telephone speech
Price
$7,000.00 (non-member fee)
per corpus of about 984 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $3,500.00.
Source
LDC catalog, Fisher English Training Speech Part 1 Speech (LDC2004S13)
A catalogue or marketplace listing
Fisher Levantine Arabic Conversational Telephone Speech: 45 hours (279 calls)
Price
$1,000.00 (non-member fee)
per corpus of about 45 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $500.00.
Source
LDC catalog, Fisher Levantine Arabic Conversational Telephone Speech (LDC2007S02)
A catalogue or marketplace listing
Fisher Spanish Speech: roughly 163 hours of Spanish telephone speech from 136 speakers
Price
$2,500.00 (non-member fee)
per corpus of about 163 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.
Source
LDC catalog, Fisher Spanish Speech (LDC2010S01)
A catalogue or marketplace listing
GALE Phase 2 Arabic Broadcast News Speech Part 1: about 165 hours of Arabic broadcast news
Price
$2,000.00 (non-member fee)
per corpus of about 165 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Source
LDC catalog, GALE Phase 2 Arabic Broadcast News Speech Part 1 (LDC2014S07)
A catalogue or marketplace listing
Global TIMIT Learner Treebank English: about 24 hours of native and non-native (L2) English read speech
Price
$750.00 (non-member fee)
per corpus of about 24 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $375.00.
Source
LDC catalog, Global TIMIT Learner Treebank English (LDC2020S09)
A catalogue or marketplace listing
HKUST Mandarin Telephone Speech Part 1: about 149 hours of Mandarin conversational telephone speech
Price
$3,000.00 (non-member fee)
per corpus of about 149 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, HKUST Mandarin Telephone Speech, Part 1 (LDC2005S15)
A catalogue or marketplace listing
IARPA Babel Zulu Language Pack: about 211 hours of Zulu conversational and scripted telephone speech
Price
$25.00 (non-member fee)
per corpus of about 211 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $25.00.
Source
LDC catalog, IARPA Babel Zulu Language Pack IARPA-babel206b-v0.1e (LDC2017S19)
A catalogue or marketplace listing
ICSI Meeting Speech: about 72 hours of English meeting speech (883 hours of multi-microphone audio)
Price
$1,900.00 (non-member fee)
per corpus of about 72 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $950.00.
Source
LDC catalog, ICSI Meeting Speech (LDC2004S02)
A catalogue or marketplace listing
Korean Telephone Conversations Speech: about 44 hours of audio (100 calls)
Price
$1,500.00 (non-member fee)
per corpus of about 44 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $750.00.
Source
LDC catalog, Korean Telephone Conversations Speech (LDC2003S03)
A catalogue or marketplace listing
King Saud University Arabic Speech Database: 590 hours of Arabic speech from 269 speakers
Price
$2,000.00 (non-member fee)
per corpus of about 590 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Source
LDC catalog, King Saud University Arabic Speech Database (LDC2014S02)
A catalogue or marketplace listing
Magic Data Chinese Mandarin Conversational Speech: about 10 hours of conversation from 60 speakers (about 60 hours of multi-device audio)
Price
$750.00 (non-member fee)
per corpus of about 10 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $375.00.
Source
LDC catalog, Magic Data Chinese Mandarin Conversational Speech (LDC2019S23)
A catalogue or marketplace listing
MATERIAL Somali-English Language Pack: about 80 hours of Somali telephone speech with transcripts and translations
Price
$250.00 (non-member fee)
per corpus of about 80 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $125.00.
Source
LDC catalog, MATERIAL Somali-English Language Pack (LDC2024S10)
A catalogue or marketplace listing
MATERIAL Swahili-English Language Pack: about 112 hours of Swahili telephone speech with transcripts and translations
Price
$250.00 (non-member fee)
per corpus of about 112 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership.
Source
LDC catalog, MATERIAL Swahili-English Language Pack (LDC2026S01)
A catalogue or marketplace listing
Mixer 6 CHiME 8 Transcribed Calls and Interviews: 80 hours of English interviews and telephone speech with transcripts
Price
$3,000.00 (non-member fee)
per corpus of about 80 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, Mixer 6 - CHiME 8 Transcribed Calls and Interviews (LDC2025S07)
A catalogue or marketplace listing
Multi-Language Conversational Telephone Speech 2011, South Asian: about 118 hours of Bengali, Hindi, Punjabi, Tamil and Urdu telephone speech
Price
$2,500.00 (non-member fee)
per corpus of about 118 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.
Source
LDC catalog, Multi-Language Conversational Telephone Speech 2011 -- South Asian (LDC2017S14)
A catalogue or marketplace listing
Multi-Channel WSJ Audio: about 100 hours of British English read speech from 45 speakers, far-field microphone arrays
Price
$1,500.00 (non-member fee)
per corpus of about 100 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $750.00.
Source
LDC catalog, Multi-Channel WSJ Audio (LDC2014S03)
A catalogue or marketplace listing
Polish Speech Database: about 280 hours of read Polish speech from 200 speakers
Price
$3,000.00 (non-member fee)
per corpus of about 280 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, Polish Speech Database (LDC2019S19)
A catalogue or marketplace listing
REMIX Telephone Collection: 320 hours of English conversational telephone speech from 358 speakers
Price
$2,500.00 (non-member fee)
per corpus of about 320 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,250.00.
Source
LDC catalog, REMIX Telephone Collection (LDC2023S09)
A catalogue or marketplace listing
Samromur Children Icelandic Speech 1.0: 131 hours of prompted speech from 3,175 children aged 4 to 17
Price
$250.00 (non-member fee)
per corpus of about 131 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $125.00.
Source
LDC catalog, Samrómur Children Icelandic Speech 1.0 (LDC2022S11)
A catalogue or marketplace listing
Mandarin-English Code-Switching in South-East Asia (SEAME): about 192 hours from 156 speakers
Price
$2,000.00 (non-member fee)
per corpus of about 192 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Source
LDC catalog, Mandarin-English Code-Switching in South-East Asia (LDC2015S04)
A catalogue or marketplace listing
Switchboard-1 Release 2: about 260 hours of US English conversational telephone speech (543 speakers)
Price
$3,000.00 (non-member fee)
per corpus of about 260 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,500.00.
Source
LDC catalog, Switchboard-1 Release 2 (LDC97S62)
A catalogue or marketplace listing
TORGO Database of Dysarthric Articulation: about 23 hours of English speech from speakers with cerebral palsy or ALS and controls
Price
$1,200.00 (non-member fee)
per corpus of about 23 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $600.00.
Source
LDC catalog, TORGO Database of Dysarthric Articulation (LDC2012S02)
A catalogue or marketplace listing
United Nations Proceedings Speech: about 8,500 hours of proceedings in the six official UN languages
Price
$5,000.00 (non-member fee)
per corpus of about 8,500 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $2,500.00.
Source
LDC catalog, United Nations Proceedings Speech (LDC2014S08)
A catalogue or marketplace listing
CSR-I (WSJ0) Complete: about 141 hours of English read speech (123 speakers reading the Wall Street Journal)
Price
$1,500.00 (non-member fee)
per corpus of about 141 hours
When
Oct 2026
What to know
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $750.00.
Source
LDC catalog, CSR-I (WSJ0) Complete (LDC93S6A)
A catalogue or marketplace listing
35,000-hour multilingual conversational speech dataset (Nexdata, 100+ languages)
Price
$20,000 / purchase (starts at)
per purchase (one-off)
When
Oct 2026
What to know
A starting list price for bulk off-the-shelf audio, well under $1 per hour if the whole set is bought; final price is quoted per scope.
Source
Datarade (Nexdata listing)
A vendor's published price guide
Arabic podcast audio, dual-channel, 6,024 hours (Infobay.AI on Opendatabay)
Price
$29,300
per dataset of 6,024 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-16.
Source
Opendatabay marketplace, 6K+ Hours Arabic Dual-Channel Podcast Audio Dataset
A catalogue or marketplace listing
English podcast audio, dual-channel, 1,165 hours (Infobay.AI on Opendatabay)
Price
$5,660
per dataset of 1,165 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-17.
Source
Opendatabay marketplace, 1K+ Hours English Dual-Channel Podcast Audio Dataset
A catalogue or marketplace listing
German call-center recordings, dual-channel, 212 hours (Infobay.AI on Opendatabay)
Price
$1,320
per dataset of 212 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-11.
Source
Opendatabay marketplace, 200+ Hours German Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Hindi call-center recordings, dual-channel, 1,587,658 hours (Infobay.AI on Opendatabay)
Price
$826,700
per dataset of 1,587,658 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-11.
Source
Opendatabay marketplace, 1.5Mn+ Hours Hindi Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Hindi podcast audio, dual-channel, 11,607 hours (Infobay.AI on Opendatabay)
Price
$56,400
per dataset of 11,607 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-17.
Source
Opendatabay marketplace, 11K+ Hours Hindi Dual-Channel Podcast Audio Dataset
A catalogue or marketplace listing
Mizo (low-resource language of north-east India) call-center recordings, dual-channel, 468 hours (Infobay.AI on Opendatabay)
Price
$2,370
per dataset of 468 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.
Source
Opendatabay marketplace, 400+ Hours Mizo Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Nepali call-center recordings, dual-channel, 229,645 hours (Infobay.AI on Opendatabay)
Price
$487,100
per dataset of 229,645 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.
Source
Opendatabay marketplace, 229K+ Hours Nepali Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Odia call-center recordings, dual-channel, 12,794 hours (Infobay.AI on Opendatabay)
Price
$62,400
per dataset of 12,794 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.
Source
Opendatabay marketplace, 12K+ Hours Odia Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Somali call-center recordings, dual-channel, 1,057 hours (Infobay.AI on Opendatabay)
Price
$5,260
per dataset of 1,057 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.
Source
Opendatabay marketplace, 1K+ Hours Somali Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Swahili call-center recordings, dual-channel, 194,331 hours (Infobay.AI on Opendatabay)
Price
$292,100
per dataset of 194,331 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.
Source
Opendatabay marketplace, 194K+ Hours Swahili Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Tamil call-center recordings, dual-channel, 15,056 hours (Infobay.AI on Opendatabay)
Price
$73,300
per dataset of 15,056 hours
When
Oct 2026
What to know
Infobay.AI's asking price on the Opendatabay marketplace (prices in USD since 1 October 2026, converted from GBP at $1 = 0.76 GBP); the listing says final pricing depends on hours, metadata and annotation, so it is a list price, not a quote. Listed 2026-07-14.
Source
Opendatabay marketplace, 15K+ Hours Tamil Dual-Channel Call Center Dataset
A catalogue or marketplace listing
Sound-effects catalogue for AI training: 350+ libraries, about 53,000 files (vehicles, machinery, ambiences, foley) from Krampfstadt Studio, on Opendatabay
Price
$78,900
per catalogue of 53k sound files
When
Oct 2026
What to know
A studio's whole sound-effects back catalogue offered for AI training. The same seller lists it on Datarade at EUR 60,000 one-off, so the asking price differs by marketplace.
Source
Opendatabay marketplace, Krampfstadt Studio 350+ SFX libraries
A catalogue or marketplace listing
Saudi Arabic full-duplex customer-service dialogue speech (268 speakers, transcribed), 268 hours (Nexdata on Opendatabay)
Price
$118,400
per dataset of 268 hours
When
Oct 2026
What to know
Nexdata's asking price on the Opendatabay marketplace for a commercial licence with instant download; Gulf Arabic customer-service dialogue is among the dearest speech on the market. Listed 2026-08-14.
Source
Opendatabay marketplace, 268 Hours Arabic Saudi Full Duplex Speech Data
A catalogue or marketplace listing
INTERSPEECH 2025 MLC-SLM challenge data: conversational speech in 11 languages, 1,500 hours, curated by Datatang (Nexdata on Opendatabay)
Price
$526,300
per dataset of 1,500 hours
When
Oct 2026
What to know
Nexdata's asking price on Opendatabay for the commercial licence of a set first released for a research challenge; Nexdata is Datatang's international brand. Listed 2026-08-14.
Source
Opendatabay marketplace, INTERSPEECH 2025 MLC SLM Challenge Data
A catalogue or marketplace listing
Tagalog full-duplex multi-channel spontaneous dialogue speech (smartphone, transcribed), 1,100 hours (Nexdata on Opendatabay)
Price
$394,700
per dataset of 1,100 hours
When
Oct 2026
What to know
Nexdata's asking price on the Opendatabay marketplace for a commercial licence with instant download. Full-duplex, speaker-separated dialogue is the premium tier of conversational speech. Listed 2026-08-14.
Source
Opendatabay marketplace, 1100 Hours Tagalog Full Duplex Multi Channel Speech Data
A catalogue or marketplace listing
US English real customer-service calls, dual-channel, with verbatim transcripts, 1,000 hours (PissedConsumer on Opendatabay)
Price
$24,500
per dataset of 1,000 hours
When
Oct 2026
What to know
Asking price for a 1,000-hour package of real (not staged) PII-redacted support calls; larger volumes are quoted on request. Listed 2026-07-30.
Source
Opendatabay marketplace, Audio Data: English (US) Conversational Speech Audio Dataset
A catalogue or marketplace listing
Hindi real customer-service calls with Hindi-English code-switching, dual-channel, with verbatim transcripts, 1,000 hours (PissedConsumer on Opendatabay)
Price
$24,500
per dataset of 1,000 hours
When
Oct 2026
What to know
Same seller and same price as its US English package, so this seller does not charge a language premium for Hindi. Listed 2026-07-30.
Source
Opendatabay marketplace, Audio Data: Hindi Conversational Speech Audio Dataset
A catalogue or marketplace listing
Human transcription of English audio, 99%+ accuracy, 12-hour turnaround (Rev)
Price
$1.99 /min
per audio minute
When
Oct 2026
What to know
Pay-as-you-go list price for human transcription, the labelling step that turns raw audio into ASR training data; subscriptions save up to 15%.
Source
Rev pricing page
A vendor's published price guide
Amharic (Ethiopia) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Amharic (Ethiopia) conversational speech dataset
A catalogue or marketplace listing
Arabic (Egyptian) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Arabic (Egyptian) conversational speech dataset
A catalogue or marketplace listing
Arabic (Modern Standard, Saudi Arabia) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$90 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Arabic (Modern Standard, Saudi Arabia) conversational speech dataset
A catalogue or marketplace listing
Bengali (Bangladesh) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Bengali (Bangladesh) conversational speech dataset
A catalogue or marketplace listing
English (South African) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$60 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, English (South African) conversational speech dataset
A catalogue or marketplace listing
English (US) conversational speech dataset, 1,500 hours of two-person conversations from 150 native speakers, transcribed, commercial licence
Price
$90 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, English (US) conversational speech dataset
A catalogue or marketplace listing
French (DR Congo) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, French (DR Congo) conversational speech dataset
A catalogue or marketplace listing
French (France) conversational speech dataset, 1,000 hours of two-person conversations from 100 native speakers, transcribed, commercial licence
Price
$95 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, French (France) conversational speech dataset
A catalogue or marketplace listing
German conversational speech dataset, 1,000 hours of two-person conversations from 50 native speakers, transcribed, commercial licence
Price
$95 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, German conversational speech dataset
A catalogue or marketplace listing
Hausa (Nigeria) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Hausa (Nigeria) conversational speech dataset
A catalogue or marketplace listing
Hindi conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Hindi conversational speech dataset
A catalogue or marketplace listing
Japanese conversational speech dataset, 500 hours of two-person conversations from 50 native speakers, transcribed, commercial licence
Price
$95 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Japanese conversational speech dataset
A catalogue or marketplace listing
Korean conversational speech dataset, 500 hours of two-person conversations from 50 native speakers, transcribed, commercial licence
Price
$95 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Korean conversational speech dataset
A catalogue or marketplace listing
Mandarin Chinese (Taiwan) conversational speech dataset, 1,500 hours of two-person conversations from 150 native speakers, transcribed, commercial licence
Price
$70 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Mandarin Chinese (Taiwan) conversational speech dataset
A catalogue or marketplace listing
Portuguese (Brazil) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Portuguese (Brazil) conversational speech dataset
A catalogue or marketplace listing
Spanish (Mexico) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Spanish (Mexico) conversational speech dataset
A catalogue or marketplace listing
Spanish (Spain) conversational speech dataset, 1,000 hours of two-person conversations from 100 native speakers, transcribed, commercial licence
Price
$85 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Spanish (Spain) conversational speech dataset
A catalogue or marketplace listing
Swahili (Kenya) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Swahili (Kenya) conversational speech dataset
A catalogue or marketplace listing
Tamil conversational speech dataset, 500 hours of two-person conversations from 50 native speakers, transcribed, commercial licence
Price
$75 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Tamil conversational speech dataset
A catalogue or marketplace listing
Urdu (Pakistan) conversational speech dataset, 1,000 hours of two-person conversations from 100 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Urdu (Pakistan) conversational speech dataset
A catalogue or marketplace listing
Yoruba (Nigeria) conversational speech dataset, 2,000 hours of two-person conversations from 200 native speakers, transcribed, commercial licence
Price
$65 per hour of audio
per hour of audio
When
Oct 2026
What to know
Off-the-shelf catalogue price per hour for a non-exclusive commercial training licence; the vendor offers volume discounts for full-dataset and multi-language purchases. Its 60-language catalogue runs $60 to $95 an hour.
Source
SpeechData.ai catalogue, Yoruba (Nigeria) conversational speech dataset
A catalogue or marketplace listing
Human transcription of audio and video, 24-hour turnaround (Speechpad)
Price
$1.30 per minute
per audio minute
When
Oct 2026
What to know
List price for standard human transcription; 6 to 12 hour turnaround is $1.80 a minute, and timestamps add $0.25 and verbatim $0.35 a minute, the two features ASR training data needs.
Source
Speechpad pricing page
A vendor's published price guide
Human-edited transcription, 99%+ accuracy, 1 to 3 day turnaround (TranscribeMe)
Price
Starting at $0.79/min
per audio minute (starting price)
When
Oct 2026
What to know
A starting price, so it marks the floor of human transcription at a vendor that also sells AI training data services.
Source
TranscribeMe pricing page
A vendor's published price guide
British English recordings evaluated for pronunciation, fluency, tone and audio quality
Price
$40 to $65 per hour
per hour worked (paid to the evaluator)
When
Oct 2026
What to know
Evaluating speech pays more than producing or transcribing it on the same board. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, English-Speaking (British) Audio Evaluator
The payer's own page
Spanish (Mexico) voice actor recording scripted and emotional audio for AI training
Price
$20 to $65 per hour
per hour worked (paid to the voice actor)
When
Oct 2026
What to know
Hourly pay for voice acting for AI; the job board is the hiring organisation for its client. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Spanish-Speaking Voice Actor for AI Training
The payer's own page
Bengali audio and video transcription and machine-transcript review
Price
$10 to $22 per hour
per hour worked (paid to the transcriber)
When
Oct 2026
What to know
Transcriber pay for Bengali. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Bengali-Speaking Audio and Video Transcription Expert
The payer's own page
Gujarati audio transcription and review for AI training
Price
$8 to $12 per hour
per hour worked (paid to the transcriber)
When
Oct 2026
What to know
Transcriber pay for an Indic language, the low end of transcription pay on the board. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Gujarati-Speaking Audio Transcription Expert
The payer's own page
Norwegian audio and video transcription and speech-segment classification
Price
$20 to $30 per hour
per hour worked (paid to the transcriber)
When
Oct 2026
What to know
Transcriber pay for a small, high-wage European language. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Norwegian-Speaking Audio Transcription Expert
The payer's own page
Punjabi audio and video transcription with speaker identification and metadata
Price
$8 to $15 per hour
per hour worked (paid to the transcriber)
When
Oct 2026
What to know
Transcriber pay for an Indic language. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Punjabi-Speaking Transcription Expert
The payer's own page
Voice owner's royalty when others generate speech with their cloned voice in the ElevenLabs Voice Library (default rate)
Price
$0.03 per 1,000 characters
per 1,000 characters generated (paid to the voice owner)
When
Oct 2026
What to know
A per-use royalty to the person whose voice is cloned, paid weekly; it counts paid users' generations only. The page says voice actors have earned a combined $5 million through the library.
Source
ElevenLabs blog, How to monetize your voice with the Voice Library
The payer's own page
Voice owner's royalty in the ElevenLabs Voice Library for HQ or rare voices (maximum requestable rate)
Price
up to $0.20 per 1,000 characters
per 1,000 characters generated (paid to the voice owner)
When
Oct 2026
What to know
The ceiling a voice owner can set; a rare accent or tone is what earns it.
Source
ElevenLabs blog, How to monetize your voice with the Voice Library
The payer's own page
Expressive voice recording for AI training, Bengali speakers in India (Outlier, via Braintrust)
Price
up to $7.50 USD per hour
per hour, ceiling for core project work (paid to the speaker)
When
Oct 2026
What to know
The same Outlier voice project pays by the speaker's language and country: $7.50 for Bengali in India, $10 Thai, $11 Vietnamese, $20 Simplified Chinese, $25 French. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Bengali Expressive Voice AI Trainer at Outlier via Braintrust
A catalogue or marketplace listing
Expressive voice recording for AI training, Simplified Chinese speakers (Outlier, via Braintrust)
Price
up to $20 USD per hour
per hour, ceiling for core project work (paid to the speaker)
When
Oct 2026
What to know
One of a set of identical Outlier voice projects priced by language; open in Australia, Canada, the UK, Malaysia, New Zealand and Singapore. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Expressive Voice AI Trainer - Simplified Chinese at Outlier via Braintrust
A catalogue or marketplace listing
Expressive voice recording for AI training, French speakers (Outlier, via Braintrust)
Price
up to $25 USD per hour
per hour, ceiling for core project work (paid to the speaker)
When
Oct 2026
What to know
The top rate in a set of identical Outlier voice projects priced by language. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, French Expressive Voice AI Trainer at Outlier via Braintrust
A catalogue or marketplace listing
Expressive voice recording for AI training, Thai speakers (Outlier, via Braintrust)
Price
up to $10 USD per hour
per hour, ceiling for core project work (paid to the speaker)
When
Oct 2026
What to know
One of a set of identical Outlier voice projects priced by language. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Thai Expressive Voice AI Trainer at Outlier via Braintrust
A catalogue or marketplace listing
Expressive voice recording for AI training, Vietnamese speakers in Vietnam (Outlier, via Braintrust)
Price
up to $11 USD per hour
per hour, ceiling for core project work (paid to the speaker)
When
Oct 2026
What to know
One of a set of identical Outlier voice projects priced by language. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Vietnamese Expressive Voice AI Trainer at Outlier via Braintrust
A catalogue or marketplace listing
Spanish voice-over recording for AI training and media, by finished hour of audio (professional home studio)
Price
USD 200 per finished hour
per finished hour of recorded audio (paid to the voice talent)
When
Sep 2026
What to know
Pay per delivered hour of edited, mastered audio, not per hour worked; professional studio-quality equipment required. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Spanish-Speaking Voice Recording Specialist
The payer's own page
Japanese audio transcription, error correction and context notes for AI training
Price
$20 to $36 per hour
per hour worked (paid to the transcriber)
When
Sep 2026
What to know
Transcriber pay for Japanese, roughly double the Indic-language rates on the same board. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Japanese-Speaking Audio Transcription Expert
The payer's own page
English phone calls to US businesses (reservations, appointments) made by Latin American contractors to help train AI systems
Price
$18 to $28 per hour
per scheduled hour (paid to the caller)
When
Sep 2026
What to know
Paid for scheduled availability in four 5-hour shifts a week, open to Argentina, Brazil, Chile, Colombia, Mexico and Peru. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, English-Speaking Customer Service Phone Support Expert
The payer's own page
Mandarin and English audio transcription and review for AI training
Price
$20 to $30 per hour equivalent
per hour equivalent, paid per task (paid to the transcriber)
When
Sep 2026
What to know
Paid per task; the hourly figure is the job's own equivalent. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Mandarin-Speaking Audio Transcription Expert
The payer's own page
Portuguese audio transcription and review, with occasional English
Price
$10 to $18 per hour equivalent
per hour equivalent, paid per task (paid to the transcriber)
When
Sep 2026
What to know
Paid per task; the hourly figure is the job's own equivalent. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Portuguese-Speaking Audio Transcription Expert
The payer's own page
Bilingual Hebrew and English audio contributor for AI training (AfterQuery)
Price
$30/hr
per hour (paid to the contributor)
When
Sep 2026
What to know
Posted about a month before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Bilingual Audio Contributor (Hebrew/English) at AfterQuery
A catalogue or marketplace listing
Paid English conversation recording study for AI training (RemoExperts, open worldwide)
Price
$20/hr
per hour (paid to the speaker)
When
Sep 2026
What to know
Posted about two weeks before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Paid English Conversation Recording Study at RemoExperts
A catalogue or marketplace listing
Tamil audio transcription and review, varied accents and dialects
Price
$10 to $24 per hour equivalent
per hour equivalent, paid per task (paid to the transcriber)
When
Aug 2026
What to know
Paid per task; the hourly figure is the job's own equivalent. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Tamil-Speaking Audio Transcription Expert
The payer's own page
Mandarin-English code-switched conversation recording (1 to 2 minute clips) by speakers in Singapore or Malaysia
Price
$2 to $5 USD per hour
per hour worked (paid to the speaker)
When
Aug 2026
What to know
The lowest speaker pay found in this run, for phone-recorded clips with no studio needed. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, Chinese- and English-Speaking Bilingual Audio Recording
The payer's own page
German speech data: recording audio samples, phonetic annotation and evaluation
Price
$30 to $65 per hour
per hour worked (paid to the contributor)
When
Aug 2026
What to know
Mixes recording with phonetic annotation and review, so it prices a skilled native speaker rather than a voice alone. OpenTrain states it is the hiring organisation for the client project.
Source
OpenTrain AI job posting, German-Speaking Speech Data Annotation Expert
The payer's own page
Voice actor recording a customer-service agent voice for cloning, African American English (Mercor)
Price
$50–$100/hr
per hour (paid to the voice actor)
When
Aug 2026
What to know
Posted about two months before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Voice Actor: CX Agent Voice Cloning (African American) at Mercor
A catalogue or marketplace listing
Voice actor recording a customer-service agent voice for cloning, Swiss German (Mercor)
Price
$50–$150/hr
per hour (paid to the voice actor)
When
Aug 2026
What to know
Posted about two months before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Voice Actor: CX Agent Voice Cloning (Swiss German) at Mercor
A catalogue or marketplace listing
Voice recorder, German and English (micro1)
Price
$30–$65/hr
per hour (paid to the speaker)
When
Aug 2026
What to know
Posted about two months before it was read. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Voice Recorder (German + English) at micro1
A catalogue or marketplace listing
Audio recording contributor with an Australian English accent (AfterQuery)
Price
$50/hr
per hour (paid to the speaker)
When
Jul 2026
What to know
A specific accent commands more than general English recording at the same buyer. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Audio Recording Contributor - Australian English Accent at AfterQuery
A catalogue or marketplace listing
French conversation partner recording natural conversations for AI training (Babel Audio, Canada and US)
Price
$17.50/hr
per hour (paid to the speaker)
When
Jul 2026
What to know
Posted about three months before it was read; Babel Audio pays per recorded hour. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, French Conversation Partner (AI Training) at Babel Audio
A catalogue or marketplace listing
Royalties paid to voice creators for licensing their cloned voices on the ElevenLabs Voice Library, cumulative to May 2026
Price
over $22 million
in total, to 10,400+ creators
When
May 2026
What to know
Royalties for the use of a voice, not a fee for training data; up from $11 million in November 2025. Over $22 million across 10,400+ creators is about $2,100 per creator on average (our arithmetic).
Source
ElevenLabs blog, $22 million earned by voice creators on ElevenLabs
The payer's own page
Hindi voice actor or actress recording for AI training (Babel Audio)
Price
$60–$150/hr
per hour (paid to the voice actor)
When
May 2026
What to know
Professional voice acting pays far above conversational recording at the same platform ($17.50 an hour). Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Hindi Voice Actor / Actress at Babel Audio
A catalogue or marketplace listing
Conversation recording roles on David AI's Babel Audio platform
Price
$24 to $28 per hour
per hour (paid to the speaker)
When
May 2026
What to know
Speaker-side pay at an audio dataset vendor that has closed six-figure contracts with large AI labs; the buyer price is not public.
Source
Contrary Research
Press report
Voice actor recording a customer-service agent voice for cloning, US English (Mercor)
Price
$50–$150/hr
per hour (paid to the voice actor)
When
Apr 2026
What to know
Voice-cloning work, where the actor's own voice becomes the product, pays the most of any speaker role found. Rate as the platform states it, carried by AITraining.jobs, which has the same owner as AIData.Broker.
Source
AITraining.jobs, Voice Actor: CX Agent Voice Cloning (USA) at Mercor
A catalogue or marketplace listing
Speech collected to a buyer's spec
Price
$150 to $400+ an hour
per hour
When
Jan 2026
What to know
What a buyer pays a vendor to record new speech: chosen speakers, languages and scenarios, with consent on file.
Source
SpeechData.ai buyer's guide
A vendor's published price guide
Recorded phone-call audio sold to AI developers (Neon Mobile app, paid to callers)
Price
$0.15 per minute, doubled when both callers use the app, max $30 per day
per minute (paid to the caller)
When
Sep 2025
What to know
This is what the app pays the person recording, not what AI buyers pay Neon; the app later went offline after a security flaw.
Source
Malwarebytes
Press report
Open African-language speech collection: 9,000 hours across 18 languages in Kenya, Nigeria and South Africa (Gates Foundation grant)
Price
$2.2 million
per programme of 9,000 hours
When
Sep 2025
What to know
What a funded collection programme cost per hour, including recruiting speakers, recording, transcription and research staff; the data was released open access, so no seller recovers it.
Source
CO/AI news, African researchers release 9,000-hour speech dataset
Press report
Defined.ai Data Access Plan, Elite tier: 3,125 hours of off-the-shelf speech data from its marketplace
Price
€500,000 for 3,125 hours
per plan of 3,125 hours
When
Sep 2024
What to know
A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.
Source
Defined.ai press release, Data Access Plan (DAP)
The payer's own page
Defined.ai Data Access Plan, Premium tier: 625 hours of off-the-shelf speech data from its marketplace
Price
€100,000 for 625 hours
per plan of 625 hours
When
Sep 2024
What to know
A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.
Source
Defined.ai press release, Data Access Plan (DAP)
The payer's own page
Defined.ai Data Access Plan, Standard tier: 315 hours of off-the-shelf speech data from its marketplace
Price
€50,000 for 315 hours
per plan of 315 hours
When
Sep 2024
What to know
A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.
Source
Defined.ai press release, Data Access Plan (DAP)
The payer's own page
Defined.ai Data Access Plan, Ultimate tier: 1,565 hours of off-the-shelf speech data from its marketplace
Price
€250,000 for 1,565 hours
per plan of 1,565 hours
When
Sep 2024
What to know
A bundle of hours from Defined.ai's off-the-shelf speech catalogue, sold as a discount on its standard rates; per hour, every tier works out at about EUR 160.
Source
Defined.ai press release, Data Access Plan (DAP)
The payer's own page
Rural Indian speakers recording speech data for Karya (supplier to Microsoft and Google)
Price
about $5 an hour
per hour (paid to the speaker)
When
Dec 2023
What to know
Worker pay, well above the Indian minimum wage; workers also keep a share of the data for royalties.
Source
Context (Thomson Reuters Foundation)
Press report
One hour of Odia (low-resource Indian language) speech data
Price
used to cost about $3-$4, now it's $40
per hour of audio
When
Dec 2023
What to know
Karya co-founder Safiya Husain's figure for what buyers now pay; a tenfold rise as big tech sought Indian-language data.
Source
Context (Thomson Reuters Foundation)
Press report
Corpus of Spontaneous Japanese (CSJ), 661 hours of mostly monologue speech with 7.52 million words of transcripts: a 2-year commercial licence for up to 10 users
Price
500,000 yen (tax not included)
per licence of the 661-hour corpus
When
Jun 2021
What to know
The corpus owner's own commercial fee schedule (last updated 1 June 2021); an unlimited-user commercial licence is 10 million yen. The same corpus costs EUR 43,125 through ELRA.
Source
NINJAL, Corpus of Spontaneous Japanese usage fees
A catalogue or marketplace listing
Corpus of Spontaneous Japanese (CSJ), 661 hours of mostly monologue speech with 7.52 million words of transcripts: an open-ended commercial licence for up to 10 users
Price
5,000,000 yen (tax not included)
per licence of the 661-hour corpus
When
Jun 2021
What to know
The corpus owner's own commercial fee schedule (last updated 1 June 2021); an unlimited-user commercial licence is 10 million yen. The same corpus costs EUR 43,125 through ELRA.
Source
NINJAL, Corpus of Spontaneous Japanese usage fees
A catalogue or marketplace listing
Off-the-shelf speech recognition training datasets (wake words, read sentences, dialogues; Chinese and foreign languages), usage licences sold by Haitian Ruisheng in 2020: 198,951 hours
Price
413 元/小时 (CNY 413 an hour, average over 198,951 hours sold)
per hour, average across all speech recognition dataset hours licensed in 2020
When
Dec 2020
What to know
Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2020, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate. The same filing splits 2020 by language: Chinese 91,455 hours at CNY 255 an hour, foreign languages 108,525 hours at CNY 674 an hour (all speech, both kinds).
Source
Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table
The payer's own page
Off-the-shelf speech synthesis (TTS) training datasets, usage licences sold by Haitian Ruisheng in 2020: 1,029 hours
Price
13,832 元/小时 (CNY 13,832 an hour, average over 1,029 hours sold)
per hour, average across all speech synthesis dataset hours licensed in 2020
When
Dec 2020
What to know
Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2020, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.
Source
Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table
The payer's own page
Off-the-shelf speech recognition training datasets (wake words, read sentences, dialogues; Chinese and foreign languages), usage licences sold by Haitian Ruisheng in 2019: 136,411 hours
Price
549 元/小时 (CNY 549 an hour, average over 136,411 hours sold)
per hour, average across all speech recognition dataset hours licensed in 2019
When
Dec 2019
What to know
Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2019, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate. The same filing splits 2019 by language: Chinese 55,957 hours at CNY 317 an hour, foreign languages 80,728 hours at CNY 793 an hour.
Source
Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table
The payer's own page
Off-the-shelf speech synthesis (TTS) training datasets, usage licences sold by Haitian Ruisheng in 2019: 274 hours
Price
24,946 元/小时 (CNY 24,946 an hour, average over 274 hours sold)
per hour, average across all speech synthesis dataset hours licensed in 2019
When
Dec 2019
What to know
Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2019, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.
Source
Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table
The payer's own page
Off-the-shelf speech recognition training datasets (wake words, read sentences, dialogues; Chinese and foreign languages), usage licences sold by Haitian Ruisheng in 2018: 115,636 hours
Price
504 元/小时 (CNY 504 an hour, average over 115,636 hours sold)
per hour, average across all speech recognition dataset hours licensed in 2018
When
Dec 2018
What to know
Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2018, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.
Source
Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table
The payer's own page
Off-the-shelf speech synthesis (TTS) training datasets, usage licences sold by Haitian Ruisheng in 2018: 96 hours
Price
20,124 元/小时 (CNY 20,124 an hour, average over 96 hours sold)
per hour, average across all speech synthesis dataset hours licensed in 2018
When
Dec 2018
What to know
Realised sales: the seller licenses the use of datasets it owns (one dataset can be licensed many times), and the prospectus reports hours sold and the average price per hour for 2018, audited for the IPO. These are pre-generative-AI speech corpora sold mostly to Chinese and foreign tech firms; read them as the price history of off-the-shelf speech licensing, not a 2026 rate.
Source
Beijing Haitian Ruisheng Science Technology (SpeechOcean), STAR Market IPO prospectus (招股说明书), 9 Aug 2021, training data products table
The payer's own page

What yours might fetch

A first answer or two gives a range. The full valuation asks the rest and shows where every figure comes from.

What is your data worth?

Kind of data

Hours of audio

Kind of speech

$5,000to$30,000

On two answers, against 2 public prices. An indication, not an offer.

Full valuation

No data upload
No account
Free

Why labs pay for it

Voice models are trained mostly on people reading aloud, which is not how anyone talks. Real conversation, with overlap, hesitation, accents and languages the big datasets miss, is what they lack.

RisingCustom collection sells at $150 to $400 an hour, and one small multilingual specialist reports seven-figure revenue (October 2026). Catalogue prices for plain read speech are falling.

Who holds it

Podcast networks and broadcasters
Call centres and support teams
Research groups
Language communities

What it looks like

Two-way conversations with natural interruptions
Support calls recorded with consent
Interview and oral-history archives
Speech in less common languages and accents

Raises the price

Two-way conversation instead of read speech
A less common language or a strong regional accent
Accurate transcripts and speaker labels
Written consent from every speaker

Lowers it

Speakers who never agreed to this use, which makes it unsellable
Poor audio quality or heavy background music
Read speech in English, which is plentiful

The rights question

Every voice is biometric data in some jurisdictions. Each speaker must have agreed to this use; a recording made for quality and training does not cover selling it.

How we take it to buyers

We take consented archives to speech buyers and vendors. Where the speech does not exist yet, we commission it to a buyer's spec through AITraining.jobs.

Who buys it

Troveo

"AI labs and startups are actively paying six figures for company data exports", and Troveo pays the holder "on every sale, not just the first".

Troveo, business data page

Protege

A revenue share per deal, by the agreed partnership terms.

Protege, data providers page

Defined.ai

No rate card. It states "$600K+ average partner revenue" and "$1M+" for five or more partners.

Defined.ai, partnerships page

Sieve

Pays after delivery is checked ("Sieve Checks them, then pays"); recording is paid per hour and no rate is published.

Sieve, partners page

Kled

Publishes nothing about price.

Kled, site

Datarade

The seller sets the price. Datarade takes 30% commission, or 20% to 15% on a paid plan of $6,000 to $12,000 a year.

Datarade, provider page

FileYield

No figures. It charges sellers 8% and buyers 6% commission.

FileYield, sell data page

Opendatabay

The seller keeps 70% of a sale up to $100, rising to 95% above $250,000 (the commission falls from 30% to 5% as the sale grows). In a brokered sale the seller gets the agreed provider price and Opendatabay sets its own price to the buyer.

Opendatabay, fees and commission schedule

Every buyer and what it publishes

Where audio comes from

The whole supply map

2,666,000
The US had 2,666,000 customer service representative jobs in 2025, and BLS projects a 5% decline for 2025 to 2035.
Reaches buyers: Mostly not reaching buyers; recorded calls held by enterprises and BPOs
US Bureau of Labor Statistics, Occupational Outlook Handbook: Customer Service Representatives, 11 Oct 2026
42593 total hours; 29295 validated hours; 295 languages
Common Voice Scripted Speech v27.0 (data cut-off 2026-09) has 42,593 total hours, 29,295 validated hours and 295 languages.
Reaches buyers: Open download (CC0)
Mozilla Common Voice, cv-dataset README, Sep 2026
14+ billion call minutes recorded annually
Five9, one contact-centre platform, says 14+ billion call minutes are recorded on it each year for its 4,000+ customers.
Reaches buyers: Not reaching AI buyers: recordings belong to the call centres' customers and are held under processor terms
Five9, About, 11 Oct 2026
3,824,033 podcasts; 195,710,248 episodes
Listen Notes counts at least 3,824,033 podcasts and 195,710,248 episodes.
Reaches buyers: Not yet reaching buyers at scale; via brokers and podcast networks
Listen Notes, Podcast Stats, 16 Oct 2026
217,795 hours
Shaip's speech catalogue lists 217,795 hours of English speech across 61 datasets, and 55 call-centre datasets.
Reaches buyers: Vendor off-the-shelf licence
Shaip, Speech Data Catalog, 11 Oct 2026
7 million podcast titles; 700,000 audiobooks
Spotify hosts 7 million podcast titles and 700,000 audiobooks.
Reaches buyers: Platform-held; rights sit with publishers and creators
Spotify Newsroom, About Spotify, 11 Oct 2026

What a buyer's counsel will ask

19 questions apply to audio, and 13 of them block a sale until settled. Most data that fails to sell fails here, not on price.

Run the rights check Read every question

Other kinds of data

Company records Source code Video and footage Images Text and archives Datasets, tasks and environments Physical-world data Expert and human work

Hold audio? Describe it and we take it to the buyers.

Offer audio I need this data

In this topic