Setup & Installation
openclaw skills install @pengling9405/telegram-multilingual-voice-reply
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
npx clawhub install telegram-multilingual-voice-reply
What This Skill Does
Handles voice messages in Telegram with local speech-to-text (MLX/Qwen3-ASR on Apple Silicon) and text-to-speech. When a user sends a voice note, it transcribes locally, generates a reply in the same language, and sends back a single Telegram voice message with a matching text caption.
When to use it
- Replying to Telegram voice notes with both voice and text in one message
- Transcribing Chinese or English voice messages locally without cloud APIs
- Following the input language automatically (Chinese in, Chinese out)
- Switching between voice-only, text-only, or combined replies on request
- Running speech recognition on Apple Silicon with zero API costs
Example Workflow
Here's how your AI assistant might use this skill in practice.
INPUT
User sends a Chinese voice note in Telegram asking about tomorrow's weather
AGENT
- 1Transcribes the voice note locally using mlx_asr.py with Qwen3-ASR
- 2Detects the language as Chinese and generates a reply in Chinese
- 3Converts the reply text to speech with mlx_tts_voice.py
- 4Sends a single Telegram voice message with the text as caption
OUTPUT
One Telegram message with a Chinese voice reply and matching text caption