// the find
travisvn/openai-edge-tts
Free, high-quality text-to-speech API endpoint to replace OpenAI, Azure, or ElevenLabs
A local HTTP server that wraps Microsoft Edge's TTS service behind an OpenAI-compatible `/v1/audio/speech` endpoint. Intended as a drop-in replacement for OpenAI TTS in tools like Open WebUI or AnythingLLM without paying per character. Free because it uses Microsoft's online service, not because it runs locally.
OpenAI API compatibility is tight enough that most clients require zero code changes — just swap the base URL. SSE streaming support is a real addition that raw edge-tts doesn't give you. Voice selection is flexible: you can use OpenAI aliases or any edge-tts voice name directly, including non-English ones. Docker one-liner gets you running in under a minute.
It's not actually free offline — it relays every request to Microsoft's servers, so you're trading one vendor dependency for another with worse terms and no SLA. The GPL-3.0 license with a carve-out requiring you to email the author for non-personal use is unusual and legally awkward. No authentication beyond a single shared API key, so don't expose this on a public IP. The codebase is tiny (five Python files), which means anything beyond the happy path — rate limiting, per-user voices, request queuing — you're writing yourself.