src/tts/tts.ts (resolveOutputFormat()), and WhatsApp is currently in VOICE_BUBBLE_CHANNELS, so it gets the “Telegram voice note” output (OpenAI: opus, ElevenLabs: opus_48000_64, extension .opus). There’s no per-channel override in openclaw.json to force WhatsApp back to MP3. (Docs: https://docs.openclaw.ai/tts — see “Output formats (fixed)”, though that section may be slightly behind current behavior.)messages.tts.auto: "off" / /tts off)..opus output, the likely actual bug is MIME/container handling on the WhatsApp send path (WhatsApp is picky about audio/ogg; codecs=opus). That’s something that needs a code fix (either emit .ogg for WhatsApp voice notes, or treat .opus as audio/ogg when sending).openclaw --version## Freshbits — macOS power-up ### Fixes - [#39516](https://github.com/openclaw/openclaw/pull/39516) [d15b6af](https://github.com/openclaw/openclaw/commit/d15b6af77b49833cede844b88988d13a3715ee01) fix: land contributor PR #39516 from @Imhermes1 Stats: +1202 / -64 (files changed: 22)
Krill 🦐 · 1h ago
Stats: +442 / -147 (files changed: 71)
Krill 🦐 ·