Docs

Every setup needs an OpenAI-compatible LLM endpoint and a JEV gate endpoint. If the gate is down or slow, the…

The Twitch adapter connects with a bot account. The setup wizard can use device login or accept an existing OAuth…

The YouTube adapter follows the active broadcast chat and skips messages that were already in the chat when it…

The Discord adapter reads a bot token from an environment variable and responds to mentions or the configured…

Kick can run in read-only mode: without a posting token, replies are voiced by the avatar but are not posted to…

The bridge can serve the built avatar and its speech feed from the address configured in avatar.listen; OBS uses…

Chat replies spoken through custom TTS and realtime speech-to-speech use separate paths. The realtime voice…

The dashboard is served by the bridge and uses a login token to access its pages.

Use this reference for the bridge sections that control rate limits, moderation, memory, proactive replies,…

Plan IDs are configuration values, not the names shown on the public pricing page. Billing is off unless…

The Docker kit runs the bridge, avatar page, and dashboard on your own server. It keeps the service ports bound…

These answers describe behavior documented by the bridge setup and operations guides.

Use the bridge logs to identify the failed path. These examples preserve the real log prefixes and message text.