Human-powered multimodal data for AI

Get high-quality text, voice, image, video, or interaction-driven feedback from our diverse, vetted, and reliable participant pool.

Build more accurate models. And iterate faster.

Our pool contains 200k+ fully vetted and verified participants, who are monitored constantly by our automated Protocol algorithm. It’s balanced, bot-free, and demographically diverse. 

Diverse, vetted, and bot-free pools

Our taskers are engaged and reliable and deliver rich, accurate annotations and data.

Reliable taskers with rich answers

We have taskers from 38+ countries who speak 80+ languages. 

Representative samples are also available for the US and UK.

Global distribution with multiple languages and dialects

Our team can help you recruit annotators who are experienced in certain areas for specialized tasks.

Recruit specialized workforces

Text-based data for AI

Collect high-quality text-based AI datasets in less time

Quickly gather rich and accurate text data from for everything from document classification to large language models and RLHF.

Filter and find the right tasker for the right job, with reliable and rich free-text responses – to deliver complete datasets in under 2 hours.

Tools & integrations

Use your favourite text-based annotation tools (or add custom-built ones).

View all integrations
Voice-generated & audio data for AI

Gather intelligent, multilingual human input for Speech AI

Improve your voice generation models with audio data from Prolific’s diverse participant pool. Easily filter, communicate with, and manage the right taskers, then use your existing (or custom) tools to gather rich, intelligent, multilingual input.

Tools and Integrations

Use your favourite text-based annotation tools (or add custom-built ones).

View all integrations
Image annotation & video data for AI

Easily capture image and video-led data for AI

Make collecting custom image annotations and video inputs smoother, with video recording enabled (unlike MTurk). Recruit, train, and re-use high-performing participants who deliver insightful annotations and high-quality video recordings.

Tools and Integrations

Use your favourite text-based annotation tools (or add custom-built ones).

View all integrations
Interaction data for AI

Collect interaction-led data for behavioural studies and game design

Prolific makes capturing interaction-led data easier by enabling simultaneous participation with simple integrations. Bonus payments let you to gamify to create more real-life scenarios, by matching participant rewards to their performance.

Tools and Integrations

Use your favourite text-based annotation tools (or add custom-built ones).

View all integrations
Features

Streamline your production of multimodal AI data

Quickly filter top-performing AI teachers with stronger language comprehension and fluency or specific knowledge or skills.
Filters

Filter taskers fast

Integrate any third-party tool or custom-built tool with a simple link. 

Or integrate, scale, and automate with our API.
Tools

Use your own tools

Save specific sets of tasker filters that reflect the needs of your organisation or project. Then easily apply them again and again to new studies.
Filter-sets

Set rules and save filter-sets

Once you’ve used a tasker you can re-contact them, save them, or train them for specific tasks.
Communication

Contact, train, and retarget taskers

Put your high-performing or especially trained participants into groups and re-target these across different tasks or studies.
Participant groups

Group high-performing taskers

If you can control the uniqueness of your tasks, then set up multi-submissions to allow a participant to take your study over and over. (Thresholds configurable).
Multi-submissions

Enable multi-submissions

Data for AI that sets the standard

Planned EU legislation means AI companies will need to prove where their data has come from. All data sourced via Prolific is fully auditable, as well as being bot-free and mitigating bias.

Fully auditable
Bias and bot-free
Demographically diverse

Collect complete datasets in less than 2 hours

Getting started is simple. With no contract or subscription – just a pay-as-you-go fee for each completed response.