A minimal, fast, and modern web chat interface for mistral.rs, supporting text, vision, and speech models with drag-and-drop image and file upload, markdown rendering, and multi-model selection.

Features

Multi-model Support: Choose from multiple loaded models. See multi-model docs for advanced configuration.
Text-to-speech Model Support: Generate speech from text prompts and download as WAV files.
Speech Input Model Support: Upload audio inputs for multimodal models.
File Upload: Upload and work with files.
Drag & Drop Image Upload: Instantly preview and send images to vision models.
Markdown Output: Responses are rendered in markdown, including code blocks.
Copy Button: One-click copying for all code snippets.
Responsive UI: Clean layout for desktop and mobile.
Keyboard Shortcuts:
- Ctrl+Enter (or Cmd+Enter on Mac) to send messages.
- Button click also sends.

Quickstart

Build the app.

Note: choose the features based on this guide.

cargo run --release --features <specify feature(s) here> --bin mistralrs-web-chat -- \
  --text-model Qwen/Qwen3-4B \
  --vision-model google/gemma-3-4b-it \
  --speech-model nari-labs/Dia-1.6B \

At least one model is required (text, vision, or speech).
Multiple --text-model, --vision-model, or --speech-model can be specified.
--port is optional (defaults to 8080).

Access the app!

Open http://localhost:8080 (or your chosen port).

Development

Frontend

Edit static/index.html and supporting JS/CSS.
Uses marked.js for markdown rendering.

Backend

See main.rs for server and model loading logic.
Models are managed via CLI arguments; hot reloading is not supported.