Linux, MacOS, and Windows clients are available for download in our GitHub release page.
(Only Windows is tested in the alpha release)
You must then configure settings in the app. Specifically, to use Voice Chat, you need to have all STT, LLM, TTS configured; to only use Agentic Chat Terminal or Code Completion, you need to configure LLM.
| Modality | Supported Provider |
|---|---|
| STT | OpenAI |
| LLM | OpenAI |
| TTS | OpenAI, ElevenLabs |
(For TTS, you need to enter a voice name or voice ID which you can find from your provider. e.g. “alloy” for OpenAI TTS1, “Maltida” for ElevenLabs.)
Click the menu at top-left corner, then click either “new file” to get an empty file or “open file” to select one from your file system.
(Make sure you have configured all STT, LLM, TTS providers and API keys)
Click the microphone icon in the bottom toolbar, and grant permission if prompted. Then start chatting using voice. Any code changes will be made in the code view once the request completes.
Click the pen icon in the bottom toolbar, and start drawing using your cursor. If you wish to restart, simply toggle off and on the pen tool again.
(Make sure you have configured LLM provider and API key)
Click the “Open Chat View” icon in the bottom toolbar. Then select your desire agent, or define your own by clicking the plus icon. Next, start chatting by typing in text in the input box.
(Make sure you have configured LLM provider and API key)
Type anything in an open file, then a suggestion would become available in grey text. Press tab key to accept changes, or keep typing to refresh new suggest.