Captions Coop Wiki

Turn speech into live captions on hardware you control. No paid transcription account, no cloud AI key.

Who this wiki is for

There are three roles, and the pages are grouped by them:

• Users — you watch or read the captions. Start at Reading captions.
• Admins (event producers) — you run a captioning session: the capture page, editor, rooms and tokens, overlays and transcripts. Start at Your first event.
• Hosters (sysadmins) — you deploy and maintain the software. Start at Choose a deployment.

See Who does what for a one-screen summary.

New here? Start with the basics

• What this stack does — a one-minute overview in plain words.
• Concepts in plain words — the words used everywhere else (relay, rooms, tokens).
• Who does what — find your role.

Running the stack

Two supported ways. Prefer the first for real deployments:

• Deploy with Co-op Cloud — recommended, for production.
• Deploy with Docker Compose — good for development and evaluation.

What you can do

• Transcribe speech in its spoken language.
• Translate speech into English at the engine, and into many other languages in each viewer's overlay (show original and translation together).
• Download the transcript of what you captured.
• Send captions to an editor for human correction before they are shown.
• Show captions to an audience through an overlay, on a phone, or in OBS.