Get started
There isn’t one “right” way to start - it depends on what you’re building. Pick the path that matches your use case:
Try in browser (fastest)
If you just want to hear it working:
- open the browser demo
- allow mic access
- speak → hear translated speech back
Quickstart (Recommended)
If you want to integrate Pinch into your own workflow or pipeline, start with the SDK.
The current SDK Quickstart demonstrates file based audio translation:
- provide an audio file (WAV)
- receive a translated audio file (WAV)
- receive a transcript of the translated text
Integrations
If you’re integrating Pinch into an existing real-time audio stack:
- LiveKit (in progress)
- Pipecat (in progress)
These integrations are being stabilized and documented.
Not sure which one to pick?
If you’re new: start with Try in browser, then Pinch Quickstart.