telegram-offline-voice

What This Skill Does

The telegram-offline-voice skill is a sophisticated local TTS (Text-to-Speech) solution designed for OpenClaw agents. It transforms plain text—including messy Markdown-formatted responses—into high-quality, natural-sounding audio files formatted specifically for Telegram's voice message standard. By leveraging the Microsoft Edge-TTS engine, it provides a cost-free, offline alternative to expensive cloud-based APIs while maintaining superior control over audio pacing and clarity. The tool features an intelligent cleaning layer that strips away technical artifacts like Markdown bolding, code blocks, and URLs, ensuring the synthesized audio sounds like a human assistant rather than a robotic text-reader. Furthermore, it handles long-form content by automatically segmenting text into logical chunks, preventing the common issue of single, unmanageable audio files, and uses UUID-based file management to ensure thread safety in complex, multi-agent workflows.

Installation

To install this skill, ensure you have an environment capable of running shell commands. First, install the necessary system dependencies, specifically FFmpeg, which is required for audio format conversion. On Debian-based systems, use: 'sudo apt update && sudo apt install ffmpeg python3-pip -y'. Next, install the skill via the OpenClaw repository using the command: 'clawhub install openclaw/skills/skills/sanwecn/telegram-offline-voice'. We highly recommend using 'uv' as your Python package manager for optimized performance. Once installed, the skill resides in your project path, ready to be triggered by your AI agent whenever a text-to-voice conversion is requested.

Use Cases

This skill is perfect for users who want to receive updates from their AI agents while driving, walking, or engaging in activities where reading a screen is inconvenient. It is ideal for automating personal newsletters, daily briefings, or reading out long articles generated by your agent. Because it works entirely locally, it is also suited for users with strict privacy requirements or those who want to build a highly responsive agent without incurring per-character costs from commercial TTS services.

Example Prompts

1. "Generate a voice summary of the latest news headlines in my feed and send it as a Telegram voice message."
2. "Read out this long technical document for me, but please keep the voice tone natural and split it into several short clips."
3. "Convert my daily task list into an audio briefing I can listen to while I drive to work."

Tips & Limitations

To maximize the quality of the output, avoid including complex ASCII art or tables, as the text-to-speech engine is optimized for narrative, human-readable text. While the automated cleaning is robust, manually refining your text prompt can yield even better results for abbreviations or technical jargon. Keep in mind that this skill relies on the Edge-TTS engine, which requires an active internet connection to initially fetch audio data from Microsoft servers, even though the processing itself is locally managed within your instance. Ensure your system's storage has sufficient capacity if you are processing very large files, as temporary audio segments are saved locally before being delivered to Telegram.