mira-voice-f5

To generate a voice reply, this local TTS skill processes your text using the F5-TTS model, converts it to an audio file, and sends it directly via the Telegram Bot API. It operates entirely offline for privacy and low latency.

F5-TTS zero-shot voice synthesis generates high-quality speech from text without requiring an internet connection. It uses a specific Taiwanese female voice model locally on your machine, eliminating external API calls and preventing potential data leaks.

Yes, you can use F5-TTS completely offline to avoid external cloud APIs. This skill generates speech directly on your machine, ensuring privacy and reducing latency by avoiding data transmission to third-party servers.

Yes, you need python3, ffmpeg, and curl installed to run local voice synthesis. These dependencies are required to process text, format the generated audio file correctly, and send the voice message via the Telegram Bot API.

You trigger a spoken response by explicitly requesting a voice reply with your desired text. The skill then synthesizes the Mandarin text with a Taiwanese accent using F5-TTS and delivers it as a Telegram voice message.

Built-in TTS tools often fail because they lack proper audio formatting and API integration. This skill avoids those non-functional tools by using F5-TTS and ffmpeg to correctly format the audio file for Telegram delivery.