2 replies

Voice memo on WhatsApp

[openclaw] I don't have Discord configured as a channel, so I can't send messages there directly. But I can draft the bug report for you to post:

---

Bug Report: Audio transcription not running for WhatsApp voice notes

Version: 2026.2.2-3
OS: Windows 10 (x64)
Channel: WhatsApp (web gateway, selfChatMode)

Config:

{
  "env": { "vars": { "OPENAI_API_KEY": "sk-proj-..." } },
  "tools": {
    "media": {
      "audio": {
        "enabled": true,
        "models": [{ "provider": "openai", "model": "gpt-4o-mini-transcribe" }]
      }
    }
  }
}

{
  "env": { "vars": { "OPENAI_API_KEY": "sk-proj-..." } },
  "tools": {
    "media": {
      "audio": {
        "enabled": true,
        "models": [{ "provider": "openai", "model": "gpt-4o-mini-transcribe" }]
      }
    }
  }
}

Expected: Voice notes should be transcribed via OpenAI Whisper before reaching the agent.

Actual: Voice notes arrive as

<media:audio>

<media:audio>

with no transcription. Logs show

web-inbound

web-inbound

→

web-auto-reply

web-auto-reply

with no transcription attempt or error. Even with

diagnostics.flags: ["*"]

diagnostics.flags: ["*"]

, no media-related logs appear.

Verified:
- API key is valid (can list models)
- Audio file exists on disk (~9KB .ogg)
- Config passes validation
- Multiple gateway restarts attempted

Logs excerpt:

web-inbound: {"body":"<media:audio>","mediaPath":"...ogg","mediaType":"audio/ogg; codecs=opus"}
web-auto-reply: {"body":"[WhatsApp ...] <media:audio>","mediaType":"audio/ogg; codecs=opus","mediaPath":"..."}

web-inbound: {"body":"<media:audio>","mediaPath":"...ogg","mediaType":"audio/ogg; codecs=opus"}
web-auto-reply: {"body":"[WhatsApp ...] <media:audio>","mediaType":"audio/ogg; codecs=opus","mediaPath":"..."}

No transcription logs between these steps.

---

You can post this in the

#help

#help

#bugs

#bugs

channel at https://discord.com/invite/clawd

Want me to help you set up Discord so I can post there directly in the future?

FotC�

Friends of the Crustacean 🦞🤝•6d ago•

2 replies

Cstjean03

Voice memo on WhatsApp

{
  "env": { "vars": { "OPENAI_API_KEY": "sk-proj-..." } },
  "tools": {
    "media": {
      "audio": {
        "enabled": true,
        "models": [{ "provider": "openai", "model": "gpt-4o-mini-transcribe" }]
      }
    }
  }
}

{
  "env": { "vars": { "OPENAI_API_KEY": "sk-proj-..." } },
  "tools": {
    "media": {
      "audio": {
        "enabled": true,
        "models": [{ "provider": "openai", "model": "gpt-4o-mini-transcribe" }]
      }
    }
  }
}

Expected: Voice notes should be transcribed via OpenAI Whisper before reaching the agent.

Actual: Voice notes arrive as

<media:audio>

<media:audio>

with no transcription. Logs show

web-inbound

web-inbound

→

web-auto-reply

web-auto-reply

with no transcription attempt or error. Even with

diagnostics.flags: ["*"]

diagnostics.flags: ["*"]

web-inbound: {"body":"<media:audio>","mediaPath":"...ogg","mediaType":"audio/ogg; codecs=opus"}
web-auto-reply: {"body":"[WhatsApp ...] <media:audio>","mediaType":"audio/ogg; codecs=opus","mediaPath":"..."}

web-inbound: {"body":"<media:audio>","mediaPath":"...ogg","mediaType":"audio/ogg; codecs=opus"}
web-auto-reply: {"body":"[WhatsApp ...] <media:audio>","mediaType":"audio/ogg; codecs=opus","mediaPath":"..."}

No transcription logs between these steps.

---

You can post this in the

#help

#help

#bugs

#bugs

channel at https://discord.com/invite/clawd

Want me to help you set up Discord so I can post there directly in the future?

Voice memo on WhatsApp

Similar Threads

Voice memo on WhatsApp

Similar Threads

Similar Threads

Similar Threads