CN8 API Documentation

Convert text to natural-sounding speech across three providers (ElevenLabs, MiniMax, OpenAI).

The TTS API exposes three providers behind one endpoint. ElevenLabs and MiniMax support 13+ languages with language + gender voice selection (and voice cloning). OpenAI offers six language-agnostic voices. Voice resolution prefers explicit voiceid, then legacy voicename (ElevenLabs), then language + gender. Per-call voice cloning is available via clonefromaudio / clonefromvideo.

• POST core-tts: text + provider + (language+gender or voiceid) → audiourl + duration_seconds
• GET core-tts-voices: per-provider catalog (elevenlabs / minimax / openai), each with capability flags
• Optional per-call voice cloning: clonefromaudio or clonefromvideo