Get real-world sounds recorded by people on their own phones in the countries you choose. Describe what to record, set the pay and approve only the audio that fits.
Real places, real noiseAny countryReview every recording
190+
Countries
13,487
Tasks completed
12,000+
Registered workers
~4 Hours
Avg time to results
Launch a sound recording task yourself and start receiving audio from real contributors, without long negotiations.
Explain what to record, where, for how long and how close to the source. Add a sample file so workers hear what a good recording sounds like.
Sounds, places or eventsSample audio for referenceClear length and distance rules
Select the countries your contributors should come from, set how many files each person sends and how many submissions you need in total.
Any number of countries in one taskSeveral files per submissionYou set the total volume
Your task becomes available to workers in the selected countries. They record the sounds around them in homes, streets, shops and vehicles.
Different phones and microphonesReal acoustic environmentsNew people in every submission
Listen to every submission, approve the audio that meets your requirements and reject the rest. Keep your dataset clean from the start.
Listen to every fileApprove or reject each submissionClean data for trainingSet up a task for the exact sounds your model needs to recognize. Contributors follow your instructions and record on their own phones.
Contributors record parks, streets, markets, stations and other places as they really sound. Train models to recognize where a recording was made.
Indoor and outdoor scenesLocal environments by countryDifferent times of dayCollect recordings of washing machines, kettles, doorbells, faucets and other household sounds. Useful for smart home devices and assistants.
Appliances and devicesReal kitchens and roomsDifferent brands and modelsGet recordings of traffic, crowds, wind, fans and office hum. Use them to make speech and audio models robust in noisy conditions.
Traffic, crowds and windSteady and changing noiseMix with your own data
Contributors record door knocks, footsteps, alarms, phone rings and other short events. Train detection models for security and smart devices.
Your own list of eventsMany repetitions per soundSafe, everyday sources onlyCollect laughing, coughing, sighing, clapping and other sounds people make without words. Used for audio event detection and wellbeing apps.
Many people, many voicesShort repeatable clipsNatural, unscripted soundsPassengers record the sound inside cars, buses, trains and metro. Train in-car assistants and models that need to work on the move.
Recorded by passengers onlyEngines, roads and announcementsDifferent vehicles by countryAsk contributors to name the sound and the place in every submission. You get audio with context, ready for labeling or training.
Your setup
Choose Audio collection, set how many files each worker sends and write instructions with what to record and what to describe.
Audio collection task typeNumber of files per submissionInstructions with a sample fileYour view
Each submission arrives with the audio files and the worker's description of the sound and location, ready to review and approve.
Audio and labels in one submissionWorker country on every submissionApprove or reject each setChoose where your sounds come from and tell contributors exactly what to record. Every submission comes from a different person.
Accept workers from all countries or pick specific ones by continent or individually. Capture how streets, homes and transport sound in each market.
Describe who should record in your instructions, such as location type, phone model or living situation. Reject submissions that do not match.
Upload a reference file that shows the length, distance and quality you expect, so contributors get it right the first time.
Each worker can complete your task only once. Set the number of submissions you need and get recordings from that many different places and devices.
You decide what counts as a good recording and review every submission before it joins your dataset.
Describe the sound, length, distance and what must not be heard, such as music or conversations.
Share an example recording so contributors know exactly what you expect.
Listen to each submission and accept only audio that meets your requirements.
Launch a separate task where other workers listen to recordings and check the labels for you.
Every recording comes from a real person who chose to take your task and knew what it was for.
Your task description explains what people will record and that the audio will be used to train AI.
Nobody is assigned to your task. Each worker decides to take it and uploads their recordings 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 task. Adjust the numbers to see what your project will cost, with the 15% platform fee included.
How long it takes a worker to complete one task.
Minimum pay is $0.10 per task. Need more recordings later? Extend the same task instead of creating a new one.
Audio data collection is gathering real-world sound recordings to train and test AI models that recognize sounds, environments and events. Good datasets include many places, devices and noise levels.
Audio data covers sounds without words: environments, appliances, noise and sound events. If you need people reading or talking, see our speech data collection page.
You set the pay for each task, starting from $0.10, plus a 15% platform fee. A campaign starts from $0.30, so you can run a small test before scaling.
You decide the length in your instructions. Sound events usually need a few seconds, ambient scenes from 30 seconds to a few minutes.
Yes. Ask contributors to name the sound, the place and the device in your instructions, and every submission will include both audio and text.
You can accept workers from all countries or choose specific continents and countries when you create the task.
Describe the format, length, distance from the source and conditions you need in your instructions. Add a sample file to show the result you expect.
State in your instructions that no conversations, music or TV may be heard. Reject any submission that breaks this rule.
Yes. Extend your existing task instead of creating a new one. Workers who already took part still cannot submit again.
If you need a special format, a review flow of your own or help with your instructions, contact us through the form below and we will set it up with you.
Need high volume, rare sounds or help writing your task? Send a message and we will set it up with you.