4 Commits

Author SHA1 Message Date
Clawdbot
4af0789ebe sync: match upstream PR #56891 at rebase 2026-03-30
Brings Gitea mirror up to date with the current state of the
openclaw/openclaw PR branch, including all fixes from Codex review:

- Namespaced directive keys (fishaudio_*/fish_* prefixes only)
- Strict latency directive validation with warnings
- Code quality cleanup, s2 model removal
- Contract and directive parsing tests
- README updated with prefixed directive docs

Source: Conan-Scott/openclaw@9787ef6e (feat/fish-audio-speech-provider)
2026-03-30 18:14:52 +11:00
Clawdbot
c23359d50d fix: lower autoSelectOrder to 15 for priority over built-in providers
Microsoft Edge TTS defaults to enabled=true with autoSelectOrder=30,
and always passes isConfigured (no API key needed). This caused auto-TTS
to route to Edge instead of fish-audio when both had order=30.

Lower to 15 to win priority over OpenAI (10), ElevenLabs (20), and
Microsoft (30) in the auto-selection fallback chain.
2026-03-29 19:06:08 +11:00
Clawdbot
ed505dcce1 fix: Opus review pass — harden before building
- Widen voice ID validation to 20-64 alphanumeric (future-proof)
- Remove hardcoded default voiceId (SJ personal clone)
- Require voiceId in isConfigured + synthesize guard with clear error
- Add model header comment explaining Fish Audio's non-standard API
- Truncate error bodies to 500 chars to prevent log pollution
- Update tests and README to match
2026-03-29 18:17:06 +11:00
Clawdbot
4842dc64a5 feat: scaffold Fish Audio speech provider plugin
- index.ts: plugin entry with definePluginEntry + registerSpeechProvider
- speech-provider.ts: full SpeechProviderPlugin implementation
  - resolveConfig from messages.tts.providers.fish-audio
  - parseDirectiveToken for voice, model, speed, latency, temperature, top_p
  - listVoices merging official + user's own voices
  - synthesize with format-aware output (opus for voice-note, mp3 otherwise)
  - stub Talk Mode (resolveTalkConfig/resolveTalkOverrides)
- tts.ts: raw fishAudioTTS() fetch + listFishAudioVoices()
  - streaming chunked → buffer, error body included in exceptions
  - parallel voice listing with graceful partial failure
- speech-provider.test.ts: voice ID validation tests
- openclaw.plugin.json: speechProviders contract
- package.json: peer dep on openclaw >=2026.3.0
2026-03-29 18:14:29 +11:00