██████╗ ███████╗██╗   ██╗██╗██████╗ ███████╗
██╔══██╗██╔════╝██║   ██║██║██╔══██╗██╔════╝
██████╔╝█████╗  ██║   ██║██║██████╔╝█████╗  
██╔══██╗██╔══╝  ╚██╗ ██╔╝██║██╔══██╗██╔══╝  
██║  ██║███████╗ ╚████╔╝ ██║██████╔╝███████╗
╚═╝  ╚═╝╚══════╝  ╚═══╝  ╚═╝╚═════╝ ╚══════╝

Mistral's open-source CLI coding assistant - Rust Edition

Revibe is a Rust port of Mistral Vibe, providing a command-line coding assistant powered by Mistral's models. It offers a conversational interface to your codebase, allowing you to use natural language to explore, modify, and interact with your projects.

• Interactive Chat: A conversational AI agent that understands your requests and breaks down complex tasks.
• Powerful Toolset: Built-in tools for file manipulation, code searching, version control, and command execution.
  • bash - Execute shell commands
  • read_file - Read file contents
  • write_file - Create or overwrite files
  • search_replace - Make precise edits to files
  • grep - Search for patterns in code
  • todo - Track tasks and progress
• MCP Support: Connect to Model Context Protocol servers for extended functionality.
• Highly Configurable: Customize models, providers, tool permissions, and UI preferences.
• Safety First: Tool execution approval and pattern-based allowlists/denylists.

# Clone the repository
git clone https://github.com/nicksenger/revibe
cd revibe

# Build and install
cargo install --path crates/revibe-cli

• Rust 1.92 or later
• A Mistral API key (get one at console.mistral.ai)

1. Run revibe for the first time to set up your API key:

   revibe --setup

2. Navigate to your project directory and start chatting:

   cd /path/to/your/project
   revibe

3. Ask questions and give instructions:

   > Find all TODO comments in the project
   > Refactor the main function to be more modular
   > Write tests for the user authentication module
   

# Start interactive session
revibe

# Start with an initial prompt
revibe "Explain this codebase"

# Start in auto-approve mode (skip tool confirmations)
revibe --auto-approve

# Start in plan mode (read-only tools only)
revibe --plan

# Run a single prompt and exit
revibe -p "List all Python files in src/"

# Pipe input
echo "What does this function do?" | revibe -p

# JSON output
revibe -p "Summarize the README" --output json

# Limit cost or turns
revibe -p "Refactor this file" --max-price 1.0 --max-turns 10

In interactive mode, use these commands:

Prefix with ! to run shell commands directly:

> !git status
> !npm test

Use @ to reference files in your prompts:

> Read @src/main.rs and explain what it does
> Update @config.toml with the new settings

Configuration is stored in ~/.revibe/config.toml:

# Active model
active_model = "devstral-2"

# Auto-compact threshold (tokens)
auto_compact_threshold = 200000

# UI theme
textual_theme = "textual-dark"

# System prompt
system_prompt_id = "cli"

# Tool configurations
[tools.bash]
permission = "ask"
allowlist = ["ls", "cat", "git status"]

# Custom providers
[[providers]]
name = "custom"
api_base = "https://api.example.com/v1"
api_key_env_var = "CUSTOM_API_KEY"

# Custom models
[[models]]
name = "custom-model"
provider = "custom"
alias = "custom"

• REVIBE_HOME - Custom configuration directory (default: ~/.revibe)
• MISTRAL_API_KEY - Mistral API key

Configure MCP servers for extended functionality:

[[mcp_servers]]
name = "fetch"
transport = "stdio"
command = "mcp-server-fetch"
args = []

[[mcp_servers]]
name = "database"
transport = "http"
url = "http://localhost:8000"

revibe/
├── crates/
│   ├── revibe-cli/      # Main CLI binary
│   ├── revibe-core/     # Core agent, config, types
│   ├── revibe-llm/      # LLM backends (Mistral, OpenAI-compatible)
│   ├── revibe-tools/    # Tool abstractions and builtins
│   └── revibe-mcp/      # MCP protocol support
├── Cargo.toml           # Workspace configuration
└── README.md

# Run tests
cargo test

# Run with logging
RUST_LOG=debug cargo run -p revibe-cli

# Check formatting
cargo fmt --check

# Run clippy
cargo clippy --all-targets

This Rust implementation aims to be a 1:1 functional port with the following differences:

• Performance: Native compilation for faster startup and execution
• Memory: Lower memory footprint
• Dependencies: Uses Rust ecosystem (tokio, reqwest, clap, ratatui)
• Distribution: Single static binary, no runtime dependencies

Copyright 2025 Devstral 2

Licensed under the Apache License, Version 2.0. See LICENSE for details.

• mistral-vibe - Original Python implementation