Guide

Video recording

Collect webcam video with automatic S3 upload. Describe the recording task in the AI editor.

Collect short webcam clips from participants. Recordings upload to S3 automatically; trial data stores a video_s3_key reference rather than raw bytes.

Preview mode does not upload video. Test recordings in a real participant session after publishing. Participants must allow camera (and microphone, if you record audio with the clip).

How it works

  1. 1

    Open your experiment in the editor

    Go to Dashboard → your project → open an experiment. Use the AI panel on the right to describe the recording task in plain language — for example what to show on screen, how long they may record, and whether to include audio.

    You do not need a special protocol prompt. Asking to record webcam video is enough for the builder to wire up camera capture and S3 upload.

  2. 2

    Review the generated code

    The AI should add a live camera preview, start/stop controls, and keep prompts, duration, and labels in config.js. Adjust those CONFIG values as needed.

  3. 3

    Publish and test with a real session

    Publish your project and run it via /take/:projectId (not preview). Allow camera access when prompted. Trial data will include video_s3_key pointing to the recording in S3.

Example ask: “Add a trial where the participant records a short webcam video introducing themselves. Keep the max length in CONFIG.”

What gets saved in trial data

  • video_s3_key — S3 path to the recording (required).
  • Any other fields you define (e.g. prompt , recording_duration_ms , custom labels in CONFIG).