Collect voice recordings from native speakers in any country. Get audio with transcripts in one task for ASR, TTS and voice AI.
Native speakers worldwideAudio and transcripts in one taskReview every recording190+
Countries
13,487
Tasks completed
12,000+
Registered workers
~4 Hours
Avg time to results
Launch a speech task yourself and get voice recordings from real people without long negotiations.
Add a script to read, a topic to talk about or a list of voice commands. Ask for a transcript with each recording if you need one.
Scripts, topics or voice commandsTranscripts on requestYour own instructions and examples
Select the countries you need, set the number of recordings and describe speaker requirements such as language, accent or age in your task.
Any number of countries in one taskSpeaker requirements in your instructionsYou set the number of recordings
Your task becomes available to workers in the selected countries. They record audio on their phones and submit it together with the transcript.
Native speakers in your target marketsRecorded on real phonesAudio and transcript in one submission
Listen to each submission, check the transcript and reject anything that does not meet your requirements. Keep a clean dataset for training.
Listen to every recordingApprove or reject each submissionPay only for what you acceptSet up a task for the exact type of recordings your model needs. Workers follow your instructions and record on their own phones.
Workers read sentences, phrases or numbers from your script. You know exactly what was said in every recording, so the data is ready for training without guesswork.
Your own sentences, phrases or numbersSame script read by many different speakersText is known in advance for every recording
Workers speak freely on a topic you set, like describing their day or giving an opinion. You get real phrasing, pauses and fillers that scripted data cannot give you.
Topics and questions you defineNatural pauses, fillers and real phrasingTranscript added to every recording
Workers repeat short commands and trigger phrases in their own voice and manner. Collect the variety your voice assistant needs to respond to anyone.
Wake words, commands and short phrasesMany repetitions from different speakersDifferent ages, genders and voices
Workers record a conversation with a friend or family member, following roles or a free topic. Both participants agree to the recording.
Role-based or free conversationsReal back-and-forth speechConsent from both participants
Collect the same phrases from speakers in different countries and regions. Your model learns to understand people beyond one standard accent.
Speakers from the countries you chooseSame phrases across different regionsNative speakers of each language
Workers record on the street, in a car, in a café or at home with background sounds. Your model learns to hear speech where people actually use it.
Street, car, café or home settingsReal background noise, not added laterLocation type noted for each recording
Ask workers to send both the audio and the text of what they said. You get ready training pairs without a separate transcription step.
Your setup
Choose Audio collection, set how many files each worker sends and add instructions with phrases to read and recording rules.
Audio collection task typeNumber of files per submissionInstructions with your scriptYour view
Each submission arrives as an audio file with the transcript attached, ready to review and approve.
Audio and text in one submissionWorker country on every submissionApprove or reject each pairChoose where your speakers come from and tell them exactly how to record. Every submission comes from a different person.
Accept workers from all countries or pick specific ones by continent or individually. Only people from the selected countries can take your task.
Describe who should record in your instructions: native language, accent, age group or gender. Reject submissions that do not match.
Upload a sample recording, a script or a short video so workers hear exactly what you expect before they start.
Each worker can complete your task only once. Set the number of recordings you need and get them from that many different speakers.
You decide what counts as a good recording and review every submission before it joins your dataset.
Describe exactly what workers must record: phrases, length, tone and recording conditions.
Upload sample audio or a short video so workers hear what a good result sounds like.
Listen to each submission and accept only recordings that meet your requirements.
Launch a separate task where other workers check recordings and transcripts for you.
Every recording comes from a real person who chose to take your task and knew what it was for. No scraped audio and no unknown sources.
Your task description explains what people will record and that it will be used to train AI. Workers read it before they accept.
Nobody is assigned to your task. Each worker decides to take it and submits their recording on their own.
Each submission comes from a specific worker in the country you selected, recorded specifically for your task.
You set the pay per recording. Here is what typical tasks cost with the 15% platform fee included.
20 recordings
Check quality and instructions on a small batch before you scale.
$4.60
$0.20 per recording
100 recordings
Phrases read aloud from your script by 100 different speakers.
$23
$0.20 per recording
300 recordings
Free speech on your topic with a transcript in every submission.
$103.50
$0.30 per recording
1,000 recordings
The same phrases from speakers across many countries and regions.
$287.50
$0.25 per recording
Prices are examples. You choose the pay rate, starting from $0.10 per task.
Speech data collection is gathering voice recordings from real people to train and test AI models such as speech recognition, text-to-speech and voice assistants. Good datasets include many different speakers, accents and recording conditions.
You set the pay for each recording, starting from $0.10, plus a 15% platform fee. A campaign starts from $0.30, so you can run a small test before scaling.
Workers can start as soon as your campaign is live. Speed depends on the countries you choose, your pay rate and how complex the task is. A small test run is the quickest way to check timing and quality.
Yes. Ask workers in your submission instructions to type exactly what they said, and each submission will include both the audio and the text.
Each worker can complete your task only once, so every submission comes from a different speaker. If you need several phrases from one person, set the number of files per submission.
You can accept workers from all countries or choose specific continents and individual countries when you create the task.
Describe the speakers you need in your task instructions, for example native Hindi speakers aged 25–40. You review every submission and can reject recordings that do not match.
Any you need. Describe the file format, recording length and conditions in your submission instructions, and workers follow them when recording.
Yes. Upload up to 30 reference files, such as a sample recording, a script or a short instructional video, so workers hear what a good result sounds like.
You listen to every submission and approve or reject it. For larger projects you can also launch a separate task where other workers check recordings and transcripts.
Yes. Extend your existing task instead of creating a new one. This keeps all recordings in one place and still lets each worker submit only once.
If you need a different file type, a special review flow or help with your instructions, contact us through the form below and we will set it up with you.
Need a custom file type, a special review flow or help with your instructions? Send a message and we will set it up with you.
