# feat: dictate into the composer with a local speech model · stienswout/t3code

[View on GitCafe](https://git.cafe/stienswout/t3code/commit/05a7252619d19fe9bfcc828cba99e4f69ee02b8e)

Repository: [stienswout/t3code](https://git.cafe/stienswout/t3code)

Visibility: public

Requested revision: 05a7252619d19fe9bfcc828cba99e4f69ee02b8e

Requested commit: 05a7252619d19fe9bfcc828cba99e4f69ee02b8e

Commit: 05a7252619d19fe9bfcc828cba99e4f69ee02b8e

Tree: f487e2fc99971565c30be78a6e6a1db15bc17430

Author: Wout Stiens

Committer: Wout Stiens

## Message

```
feat: dictate into the composer with a local speech model

Adds speech-to-text dictation to the web and desktop composer. Audio
streams to an environment that runs an opt-in local model (sherpa-onnx
with Parakeet v3, Whisper turbo, or Moonshine base), so remote browsers
can dictate through the machine they are connected to.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

```

## Parents

- [4d73e1d5cdd32dbfb470e9422421cc40475d5e99](https://git.cafe/stienswout/t3code/commit/4d73e1d5cdd32dbfb470e9422421cc40475d5e99?format=markdown)

[Source at this commit](https://git.cafe/stienswout/t3code/tree/05a7252619d19fe9bfcc828cba99e4f69ee02b8e?format=markdown)