transcribe

name	transcribe
description	Transcribe audio files to text using SenseVoice offline STT. Use when the user wants to convert speech/audio to text, transcribe voice memos, or process audio files. Supports WAV, MP3, M4A, OGG, FLAC, AAC, OPUS in Chinese, English, Japanese, Korean, and Cantonese.
disable-model-invocation	false
allowed-tools	Bash(sensevoice-cli *), Bash(~/.openclaw/bin/sensevoice-cli *), Bash(*/sensevoice-cli *), Read

Audio Transcription with SenseVoice

Transcribe audio files using the local sensevoice-cli binary.

Binary locations (try in order)

1. ~/.openclaw/bin/sensevoice-cli
2. ./rust/target/release/sensevoice-cli

Usage

sensevoice-cli <audio_file_path>

Output is plain text transcription to stdout.

Steps

1. Locate the audio file path (ask user if not provided)
2. Find the sensevoice-cli binary (check both paths above)
3. Run: <binary_path> <audio_file_path>
4. Return the transcription text to the user
5. For multiple files, process sequentially and combine results

Supported formats

WAV, MP3, M4A, OGG, FLAC, AAC, OPUS

If binary not found

• Build from source: cd rust && cargo build --release
• Or install: curl -fsSL https://raw.githubusercontent.com/xxtluuu/openclaw-sensevoice/main/install.sh | bash

If model not found

Model should be at: ~/.openclaw/models/sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/ Download via install.sh or /sensevoice setup command.