This app uses speech recognition to transcribe your voice and translate it in real-time.
Default translation uses Google Cloud Translation (API key optional). It also supports OpenAI-compatible and Anthropic endpoints via URL params tprovider, turl, tmodel, and tkey.
There is non-premium version for free of course also, but it's quite a bit more limited. Check out more options and details on GitHub.
To use this app, accept microphone permissions on page load and start speaking. Then enable Translation Output. The overlay link mirrors translated output.
Important: Only open one instance of this tool at a time to avoid microphone access conflicts.
Please note that this app uses your default microphone as the audio input source. You sometimes can change the default audio source via the browser's setting, but you can also change it at your system level by changing the default recording device. You can also change audio sources by using a Virtual Audio Cable, such as this one. Using it, it becomes possible to select other sources, including microphones, speakers, and other applications.
The overlay page supports several URL parameters for customization:
clear=1 - Clear existing captions when new ones arriveshowtime=5000 - Set timeout for captions in millisecondshtml=1 - Allow HTML in captions (use with caution)speech=1 - Enable text-to-speech for captions