Documentation
¶
Overview ¶
Command voice-fish is a headless jargo voice backend.
It uses Fish Audio text-to-speech (set FISH_REFERENCE_ID), with Deepgram STT and an Anthropic LLM.
jargo is a backend framework, so this is a server only — it exposes the WebRTC signaling endpoint POST /offer and no web UI. Point a browser client at it (the nextjs-voicebot example in gojargo/jargo-client-react, with NEXT_PUBLIC_JARGO_URL=http://localhost:8080).
DEEPGRAM_API_KEY=… ANTHROPIC_API_KEY=… FISH_API_KEY=… FISH_REFERENCE_ID=… go run ./examples/voice/fish
Click to show internal directories.
Click to hide internal directories.