Use this command to install Mistral Vibe (ACP mode):
winget install --id=MistralAI.MistralVibe.ACP -e
Mistral Vibe is a command-line coding assistant powered by Mistral's models, designed to enhance developers' productivity through natural language interaction. It enables users to explore, modify, and interact with their codebases using conversational commands, offering a seamless way to manage projects.
Key Features:
Interactive Chat: Engage in conversations with an AI agent that understands your requests and breaks down complex tasks into manageable steps.
Powerful Toolset: Utilize tools for file manipulation (read/write/search), shell command execution, code searching, version control, and todo list management directly from the chat interface.
Project-Aware Context: Automatically scans project structure and Git status to provide relevant context, improving AI understanding of your work.
Advanced CLI Experience: Features autocompletion for commands and file paths, persistent history, and customizable themes for a smooth workflow.
Highly Configurable: Customize models, tools, and UI preferences via a simple config.toml file, allowing tailored setups for specific tasks or projects.
Audience & Benefit:
Ideal for developers working on coding projects, Mistral Vibe streamlines code management by enabling natural language interactions with your project. It allows you to accomplish tasks more efficiently while supporting both UNIX environments and Windows through winget installation. Whether you're refactoring code, searching for patterns, or managing todo items, Mistral Vibe enhances productivity by integrating AI-driven assistance into your development workflow.
Mistral Vibe is a command-line coding assistant powered by Mistral's models. It provides a conversational interface to your codebase, allowing you to use natural language to explore, modify, and interact with your projects through a powerful set of tools.
> [!WARNING]
> Mistral Vibe works on Windows, but we officially support and target UNIX environments.
Interactive Chat: A conversational AI agent that understands your requests and breaks down complex tasks.
Powerful Toolset: A suite of tools for file manipulation, code searching, version control, and command execution, right from the chat prompt.
Read, write, and patch files (read_file, write_file, search_replace).
Execute shell commands in a stateful terminal (bash).
Recursively search code with grep (with ripgrep support).
Manage a todo list to track the agent's work.
Ask interactive questions to gather user input (ask_user_question).
Delegate tasks to subagents for parallel work (task).
Project-Aware Context: Vibe automatically scans your project's file structure and Git status to provide relevant context to the agent, improving its understanding of your codebase.
Advanced CLI Experience: Built with modern libraries for a smooth and efficient workflow.
Autocompletion for slash commands (/) and file paths (@).
Persistent command history.
Beautiful Themes.
Highly Configurable: Customize models, providers, tool permissions, and UI preferences through a simple config.toml file.
Safety First: Features tool execution approval.
Multiple Built-in Agents: Choose from different agent profiles tailored for specific workflows.
Built-in Agents
Vibe comes with several built-in agent profiles, each designed for different use cases:
default: Standard agent that requires approval for tool executions. Best for general use.
plan: Read-only agent for exploration and planning. Auto-approves safe tools like grep and read_file.
accept-edits: Auto-approves file edits only (write_file, search_replace). Useful for code refactoring.
auto-approve: Auto-approves all tool executions. Use with caution.
Use the --agent flag to select a different agent:
vibe --agent plan
Subagents and Task Delegation
Vibe supports subagents for delegating tasks. Subagents run independently and can perform specialized work without user interaction, preventing the context from being overloaded.
The task tool allows the agent to delegate work to subagents:
> Can you explore the codebase structure while I work on something else?
🤖 I'll use the task tool to delegate this to the explore subagent.
> task(task="Analyze the project structure and architecture", agent="explore")
Create custom subagents by adding agent_type = "subagent" to your agent configuration. Vibe comes with a built-in subagent called explore, a read-only subagent for codebase exploration used internally for delegation.
Interactive User Questions
The ask_user_question tool allows the agent to ask you clarifying questions during its work. This enables more interactive and collaborative workflows.
> Can you help me refactor this function?
🤖 I need to understand your requirements better before proceeding.
> ask_user_question(questions=[{
"question": "What's the main goal of this refactoring?",
"options": [
{"label": "Performance", "description": "Make it run faster"},
{"label": "Readability", "description": "Make it easier to understand"},
{"label": "Maintainability", "description": "Make it easier to modify"}
]
}])
The agent can ask multiple questions at once, displayed as tabs. Each question supports 2-4 options plus an automatic "Other" option for free text responses.
Terminal Requirements
Vibe's interactive interface requires a modern terminal emulator. Recommended terminal emulators include:
WezTerm (cross-platform)
Alacritty (cross-platform)
Ghostty (Linux and macOS)
Kitty (Linux and macOS)
Most modern terminals should work, but older or minimal terminal emulators may have display issues.
Quick Start
Navigate to your project's root directory:
cd /path/to/your/project
Run Vibe:
vibe
If this is your first time running Vibe, it will:
Create a default configuration file at ~/.vibe/config.toml
Prompt you to enter your API key if it's not already configured
Save your API key to ~/.vibe/.env for future use
Alternatively, you can configure your API key separately using vibe --setup.
Start interacting with the agent!
> Can you find all instances of the word "TODO" in the project?
🤖 The user wants to find all instances of "TODO". The `grep` tool is perfect for this. I will use it to search the current directory.
> grep(pattern="TODO", path=".")
... (grep tool output) ...
🤖 I found the following "TODO" comments in your project.
Usage
Interactive Mode
Simply run vibe to enter the interactive chat loop.
Multi-line Input: Press Ctrl+J or Shift+Enter for select terminals to insert a newline.
File Paths: Reference files in your prompt using the @ symbol for smart autocompletion (e.g., > Read the file @src/agent.py).
Shell Commands: Prefix any command with ! to execute it directly in your shell, bypassing the agent (e.g., > !ls -l).
External Editor: Press Ctrl+G to edit your current input in an external editor.
Tool Output Toggle: Press Ctrl+O to toggle the tool output view.
Todo View Toggle: Press Ctrl+T to toggle the todo list view.
Auto-Approve Toggle: Press Shift+Tab to toggle auto-approve mode on/off.
You can start Vibe with a prompt using the following command:
vibe "Refactor the main function in cli/main.py to be more modular."
Note: The --auto-approve flag automatically approves all tool executions without prompting. In interactive mode, you can also toggle auto-approve on/off using Shift+Tab.
Trust Folder System
Vibe includes a trust folder system to ensure you only run the agent in directories you trust. When you first run Vibe in a new directory which contains a .vibe subfolder, it may ask you to confirm whether you trust the folder.
Trusted folders are remembered for future sessions. You can manage trusted folders through its configuration file ~/.vibe/trusted_folders.toml.
This safety feature helps prevent accidental execution in sensitive directories.
Programmatic Mode
You can run Vibe non-interactively by piping input or using the --prompt flag. This is useful for scripting.
vibe --prompt "Refactor the main function in cli/main.py to be more modular."
By default, it uses auto-approve mode.
Programmatic Mode Options
When using --prompt, you can specify additional options:
--max-turns N: Limit the maximum number of assistant turns. The session will stop after N turns.
--max-price DOLLARS: Set a maximum cost limit in dollars. The session will be interrupted if the cost exceeds this limit.
--enabled-tools TOOL: Enable specific tools. In programmatic mode, this disables all other tools. Can be specified multiple times. Supports exact names, glob patterns (e.g., bash*), or regex with re: prefix (e.g., re:^serena_.*$).
Use slash commands for meta-actions and configuration changes during a session.
Built-in Slash Commands
Vibe provides several built-in slash commands. Use slash commands by typing them in the input box:
> /help
Custom Slash Commands via Skills
You can define your own slash commands through the skills system. Skills are reusable components that extend Vibe's functionality.
To create a custom slash command:
Create a skill directory with a SKILL.md file
Set user-invocable = true in the skill metadata
Define the command logic in your skill
Example skill metadata:
---
name: my-skill
description: My custom skill with slash commands
user-invocable: true
---
Custom slash commands appear in the autocompletion menu alongside built-in commands.
Skills System
Vibe's skills system allows you to extend functionality through reusable components. Skills can add new tools, slash commands, and specialized behaviors.
Local project skills: .vibe/skills/ in your project
Custom paths: Configured in config.toml
skill_paths = ["/path/to/custom/skills"]
Managing Skills
Enable or disable skills using patterns in your configuration:
# Enable specific skills
enabled_skills = ["code-review", "test-*"]
# Disable specific skills
disabled_skills = ["experimental-*"]
Skills support the same pattern matching as tools (exact names, glob patterns, and regex).
Configuration
Configuration File Location
Vibe is configured via a config.toml file. It looks for this file first in ./.vibe/config.toml and then falls back to ~/.vibe/config.toml.
API Key Configuration
To use Vibe, you'll need a Mistral API key. You can obtain one by signing up at https://console.mistral.ai.
You can configure your API key using vibe --setup, or through one of the methods below.
Vibe supports multiple ways to configure your API keys:
Interactive Setup (Recommended for first-time users): When you run Vibe for the first time or if your API key is missing, Vibe will prompt you to enter it. The key will be securely saved to ~/.vibe/.env for future sessions.
Environment Variables: Set your API key as an environment variable:
export MISTRAL_API_KEY="your_mistral_api_key"
.env File: Create a .env file in ~/.vibe/ and add your API keys:
MISTRAL_API_KEY=your_mistral_api_key
Vibe automatically loads API keys from ~/.vibe/.env on startup. Environment variables take precedence over the .env file if both are set.
Note: The .env file is specifically for API keys and other provider credentials. General Vibe configuration should be done in config.toml.
Custom System Prompts
You can create custom system prompts to replace the default one (prompts/cli.md). Create a markdown file in the ~/.vibe/prompts/ directory with your custom prompt content.
To use a custom system prompt, set the system_prompt_id in your configuration to match the filename (without the .md extension):
# Use a custom system prompt
system_prompt_id = "my_custom_prompt"
This will load the prompt from ~/.vibe/prompts/my_custom_prompt.md.
Custom Agent Configurations
You can create custom agent configurations for specific use cases (e.g., red-teaming, specialized tasks) by adding agent-specific TOML files in the ~/.vibe/agents/ directory.
To use a custom agent, run Vibe with the --agent flag:
vibe --agent my_custom_agent
Vibe will look for a file named my_custom_agent.toml in the agents directory and apply its configuration.
Example custom agent configuration (~/.vibe/agents/redteam.toml):
# Custom agent configuration for red-teaming
active_model = "devstral-2"
system_prompt_id = "redteam"
# Disable some tools for this agent
disabled_tools = ["search_replace", "write_file"]
# Override tool permissions for this agent
[tools.bash]
permission = "always"
[tools.read_file]
permission = "always"
Note: This implies that you have set up a redteam prompt named ~/.vibe/prompts/redteam.md.
Tool Management
Enable/Disable Tools with Patterns
You can control which tools are active using enabled_tools and disabled_tools.
These fields support exact names, glob patterns, and regular expressions.
Examples:
# Only enable tools that start with "serena_" (glob)
enabled_tools = ["serena_*"]
# Regex (prefix with re:) — matches full tool name (case-insensitive)
enabled_tools = ["re:^serena_.*$"]
# Disable a group with glob; everything else stays enabled
disabled_tools = ["mcp_*", "grep"]
Notes:
MCP tool names use underscores, e.g., serena_list not serena.list.
Regex patterns are matched against the full tool name using fullmatch.
MCP Server Configuration
You can configure MCP (Model Context Protocol) servers to extend Vibe's capabilities. Add MCP server configurations under the mcp_servers section:
# Example MCP server configurations
[[mcp_servers]]
name = "my_http_server"
transport = "http"
url = "http://localhost:8000"
headers = { "Authorization" = "Bearer my_token" }
api_key_env = "MY_API_KEY_ENV_VAR"
api_key_header = "Authorization"
api_key_format = "Bearer {token}"
[[mcp_servers]]
name = "my_streamable_server"
transport = "streamable-http"
url = "http://localhost:8001"
headers = { "X-API-Key" = "my_api_key" }
[[mcp_servers]]
name = "fetch_server"
transport = "stdio"
command = "uvx"
args = ["mcp-server-fetch"]
env = { "DEBUG" = "1", "LOG_LEVEL" = "info" }
Supported transports:
http: Standard HTTP transport
streamable-http: HTTP transport with streaming support
stdio: Standard input/output transport (for local processes)
Key fields:
name: A short alias for the server (used in tool names)
transport: The transport type
url: Base URL for HTTP transports
headers: Additional HTTP headers
api_key_env: Environment variable containing the API key
command: Command to run for stdio transport
args: Additional arguments for stdio transport
startup_timeout_sec: Timeout in seconds for the server to start and initialize (default 10s)
tool_timeout_sec: Timeout in seconds for tool execution (default 60s)
env: Environment variables to set for the MCP server of transport type stdio
MCP tools are named using the pattern {server_name}_{tool_name} and can be configured with permissions like built-in tools:
# Configure permissions for specific MCP tools
[tools.fetch_server_get]
permission = "always"
[tools.my_http_server_query]
permission = "ask"
MCP server configurations support additional features:
Environment variables: Set environment variables for MCP servers
Custom timeouts: Configure startup and tool execution timeouts
CONTRIBUTING - Guidelines for feedback and bug reports
License
Copyright 2025 Mistral AI
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the LICENSE file for the full license text.
