ask-talk

Text To Speech

Server-side TTS proxy. Calls OpenAI audio/speech endpoint and streams audio/mpeg back to the client. The OpenAI API key is NEVER sent to the frontend.

Falls back gracefully: if OPENAI_API_KEY is not set, returns 503 so the client can fall back to browser SpeechSynthesis.

Metered since 2026-09-14. This was an optional-auth endpoint with no rate limit, no quota and no ledger row — an anonymous caller could drive OpenAI spend indefinitely (audit §6.8). Now: the same per-IP window as /api/ask/query (20/h signed-in, 5/h anonymous, fail-open on Redis), and a ledger row per call so TTS shows up in api_call_log like everything else. Guests keep the browser SpeechSynthesis fallback on 429.

post/api/talk/tts

Request

  • Base URL: https://api.activitia.net (the document declares no server; this is the origin it was published from)
  • URL: https://api.activitia.net/api/talk/tts
  • Auth: none declared

Headers

authorizationstring

Request body

textstring required
voicestring

Response

Successful Response

unknown required

Changes

No recorded changes to this operation across all 1 revision of this API.