#Hey — I’m trying to get Telegram DM voice-note transcription working on a Mac mini (Apple Silicon) w

1 messages · Page 1 of 1 (latest)

hollow python
#

Hey — I’m trying to get Telegram DM voice-note transcription working on a Mac mini (Apple Silicon) with OpenClaw 2026.3.28.

I already have:

  • local MLX Whisper daemon working on 127.0.0.1:8787
  • ffmpeg installed
  • model downloaded
  • tools.media.audio configured with a CLI model pointing at a local transcribe.sh
  • gateway restarts cleanly

But Telegram voice notes still don’t transcribe, and the bot says it still can’t transcribe audio yet.

Is tools.media.audio.models actually the right hook for inbound Telegram DMs in 2026.3.28, or is there another Telegram-specific config path needed?