Bug type
Regression (worked before, now fails)
Beta release blocker
No
Summary
TTS audio is delivered as MP3 file attachment instead of clickable voice message in Feishu.
Steps to reproduce
- Enable TTS in OpenClaw (messages.tts.auto: always, provider: microsoft)
- Send any message via Feishu channel
- The agent replies with text AND an MP3 audio file attachment
- The audio appears as a downloadable file, not a clickable voice message
Expected behavior
TTS audio should be sent as a Feishu voice message (clickable audio bubble), not as an MP3 file attachment. According to the TTS documentation, Feishu should receive Opus voice messages.
Actual behavior
The TTS audio is delivered as an MP3 file attachment in Feishu chat. Users must download the file and play it manually, rather than tapping to play like native Feishu voice messages.
OpenClaw version
2026.3.28
Operating system
Windows 11
Install method
npm global
Model
minimax/MiniMax-M2.7
Provider / routing chain
openclaw -> feishu (channel) -> minimax (model provider)
Additional provider/model setup details
No response
Logs, screenshots, and evidence
Impact and severity
Affected: Feishu channel users with TTS enabled
Severity: Annoying (degrades user experience)
Frequency: Always (100% of TTS replies)
Consequence: Users receive MP3 files instead of voice messages, requiring manual download and play instead of tap-to-play
Additional information
The TTS documentation states that Feishu should send Opus voice messages. The Feishu plugin has both sendVoice and sendFile APIs available, but TTS audio appears to be routed through sendFile instead of sendVoice. The TTS provider in use is Microsoft Edge TTS (microsoft).
Bug type
Regression (worked before, now fails)
Beta release blocker
No
Summary
TTS audio is delivered as MP3 file attachment instead of clickable voice message in Feishu.
Steps to reproduce
Expected behavior
TTS audio should be sent as a Feishu voice message (clickable audio bubble), not as an MP3 file attachment. According to the TTS documentation, Feishu should receive Opus voice messages.
Actual behavior
The TTS audio is delivered as an MP3 file attachment in Feishu chat. Users must download the file and play it manually, rather than tapping to play like native Feishu voice messages.
OpenClaw version
2026.3.28
Operating system
Windows 11
Install method
npm global
Model
minimax/MiniMax-M2.7
Provider / routing chain
openclaw -> feishu (channel) -> minimax (model provider)
Additional provider/model setup details
No response
Logs, screenshots, and evidence
Impact and severity
Affected: Feishu channel users with TTS enabled
Severity: Annoying (degrades user experience)
Frequency: Always (100% of TTS replies)
Consequence: Users receive MP3 files instead of voice messages, requiring manual download and play instead of tap-to-play
Additional information
The TTS documentation states that Feishu should send Opus voice messages. The Feishu plugin has both sendVoice and sendFile APIs available, but TTS audio appears to be routed through sendFile instead of sendVoice. The TTS provider in use is Microsoft Edge TTS (microsoft).