Obsidian Mistral TTS

Text-to-speech plugin for Obsidian with two engines: Mistral Voxtral (cloud, best quality) and system voices (local, instant).

Features

• Two TTS engines -- switch between Mistral cloud TTS and your OS system voices
• Read notes aloud -- entire note or selected text
• Streaming playback -- starts playing in ~0.8s using Web Audio API (Mistral)
• Voice cloning -- clone any voice from a 2-3 second audio sample (Mistral)
• Voice preview -- listen to any voice before selecting it
• Save audio to vault -- next to the note or in a dedicated folder
• Markdown-aware -- strips frontmatter, headings, code blocks, and formatting before speaking
• Multiple entry points -- command palette, right-click menus, ribbon icon, status bar controls

Installation

From Obsidian Community Plugins

1. Open Obsidian Settings > Community Plugins > Browse
2. Search for "Mistral TTS"
3. Install and enable

Manual Installation

1. Download main.js, manifest.json, and styles.css from the latest release
2. Create a folder mistral-tts in your vault's .obsidian/plugins/ directory
3. Copy the three files into that folder
4. Restart Obsidian and enable "Mistral TTS" in Settings > Community Plugins

Setup

Mistral Voxtral (cloud)

1. Get an API key from console.mistral.ai
2. Open plugin settings, paste your API key
3. Create a voice by uploading a short audio sample (2-3 seconds), or use a pre-existing one
4. Select text and use "Read aloud" from the right-click menu or command palette

System Voices (local)

1. Open plugin settings, switch the TTS engine dropdown to "System voices"
2. Pick a voice from the dropdown (uses your OS built-in voices)
3. No API key needed, works offline

Commands

Command	Description
Read selection aloud	Speaks the currently selected text
Read entire note aloud	Speaks the full active note
Pause / resume playback	Toggle pause
Stop playback	Stop and reset

Audio Formats (Mistral)

Format	Use case
MP3	General use (default)
WAV	Highest quality
FLAC	Lossless compression
Opus	Low bitrate

Streaming mode uses PCM for lowest latency (~0.8s to first audio).

Supported Languages

English, French, Spanish, Portuguese, Italian, Dutch, German, Hindi, Arabic.

Development

git clone https://github.com/mcmespinaa/obsidian-mistral-tts.git
cd obsidian-mistral-tts
npm install
npm run dev    # watch mode
npm run build  # production build

Support

If you find this plugin useful, you can buy me a coffee.

License

MIT