Fish Audio speech plugin

Official OpenClaw speech provider for Fish Audio's hosted S2.1 and Drama 3 APIs.

Install it with:

openclaw plugins install @openclaw/fish-audio-speech

The plugin id is fish-audio-speech; the speech provider and TTS config id remain fish-audio. Configure tts.provider: "fish-audio" and set FISH_API_KEY. The provider supports buffered audio, HTTP-streamed playback, native Opus voice notes, 8 kHz PCM telephony, and Fish Audio voice discovery.

Set tts.providers.fish-audio.model: "drama-3-preview" to opt in to Drama 3 voice direction. The default remains s2.1-pro. Natural-language directions such as [calm, reassuring] and paired tags such as <whisper>...</whisper> are passed through in the spoken text.

See Fish Audio for setup, models, voice selection, expressive tags, and local macOS MLX usage.