vtcode

VT Code is an open-source coding agent, with semantic code intelligence via Tree-sitter. Supports multiple LLM providers with automatic failover and efficient context management.

Installation

Native Installer (Recommended) - No dependencies, instant setup:

macOS & Linux:

curl -fsSL https://raw.githubusercontent.com/vinhnx/vtcode/main/scripts/install.sh | bash

Windows (PowerShell):

irm https://raw.githubusercontent.com/vinhnx/vtcode/main/scripts/install.ps1 | iex

Alternative Installation Methods:

# Cargo (Rust package manager)
cargo install vtcode

# Homebrew (macOS/Linux)
brew install vinhnx/tap/vtcode

See Installation Guide and Native Installer Guide for more options and troubleshooting.

Usage

# Set your API key
export OPENAI_API_KEY="sk-..."

# Launch VT Code
vtcode

Supported Providers

VT Code works with OpenAI, Anthropic, Google Gemini, xAI, DeepSeek, OpenRouter, Z.AI, Moonshot AI, MiniMax, HuggingFace Inference Providers, Ollama (local & cloud), and LM Studio (local).

Set the corresponding environment variable for your provider (see Installation Guide for all options).

Agent Protocols

VT Code supports multiple agent protocols for integration:

Agent Client Protocol (ACP)

VT Code can integrate with code editors like Zed. To configure ACP, refer to the ACP docs.

Agent2Agent (A2A) Protocol

VT Code implements the Agent2Agent (A2A) Protocol, enabling communication and interoperability between AI agents. The A2A protocol supports:

• Agent Discovery: Via Agent Cards at /.well-known/agent-card.json
• Task Lifecycle Management: States like submitted, working, completed, failed
• Real-time Streaming: Via Server-Sent Events (SSE)
• Rich Content Types: Text, file, and structured data parts
• Push Notifications: Webhook-based async updates
• JSON-RPC 2.0: Over HTTP(S) for interoperability

For more information about A2A protocol support, see A2A Protocol Documentation.

Anthropic API Compatibility

VT Code provides compatibility with the Anthropic Messages API to help connect existing applications to VT Code, including tools like Claude Code.

The server supports:

• Messages endpoint at /v1/messages
• Streaming responses
• Tool calling
• Vision (image) support
• Multi-turn conversations
• System prompts
• All major Anthropic API features

For more information about Anthropic API compatibility, see Anthropic API Documentation.

Configuration

VT Code supports a rich set of configuration options, with preferences stored in vtcode.toml. Key configuration features include:

• Lifecycle Hooks: Execute shell commands in response to agent events - see Lifecycle Hooks Guide
• Tool Policies: Control which tools are allowed, prompted, or denied
• Security Settings: Configure human-in-the-loop approval and workspace boundaries
• Performance Tuning: Adjust context limits, timeouts, and caching behavior

For full configuration options, see Configuration.

CLI Design Principles

VT Code follows the Command Line Interface Guidelines and standard UNIX principles to ensure a robust and pipe-friendly experience:

• Output Streams: Primary output (e.g., code from ask, events from exec) is sent to stdout. All logs, metadata, reasoning traces, and interaction prompts are sent to stderr. This allows you to safely pipe vtcode output to other tools.

  # Only the generated code is saved to file; metadata remains on screen
  vtcode ask "factorial function in rust" > factorial.rs
  

• Standard Arguments: Adheres to GNU/POSIX argument parsing standards using clap, supporting standard flags like --help, --version, and -- separator for positional arguments.

Key Features

• Security First: Multi-layered security model with execution policy, tool policies, and argument injection protection
• Multi-Provider AI: OpenAI, Anthropic, xAI, DeepSeek, Gemini, Z.AI, Moonshot AI, OpenRouter, MiniMax, Ollama (local)
• Code Intelligence: Tree-sitter powered code navigation with go-to-definition, find-references, hover, and symbol search (Rust, Python, JavaScript/TypeScript, Go, Java, Swift, Bash)
• Smart Tools: Built-in code analysis, file operations, terminal commands, and refactoring
• Agent Skills: Comprehensive skills location system with multi-location support and precedence handling, aligned with the open Agent Skills standard (docs)
• Subagents: Delegate tasks to specialized agents (explore/plan/general/code-reviewer/debugger) via spawn_subagent; see docs/subagents/SUBAGENTS.md
• Editor Integration: Native support for Zed IDE via Agent Client Protocol (ACP)
• Lifecycle Hooks: Execute custom shell commands in response to agent events for context enrichment, policy enforcement, and automation (docs)
• Context Management: Advanced token budget tracking and context curation
• TUI Interface: Rich terminal user interface with real-time streaming

Security & Safety

VT Code implements a defense-in-depth security model to protect against prompt injection and argument injection attacks:

• Execution Policy: Command allowlist with per-command argument validation
• Workspace Isolation: All operations confined to workspace boundaries
• OS-Native Sandboxing: macOS Seatbelt and Linux Landlock + seccomp for kernel-enforced isolation
• Tool Policies: Configurable allow/deny/prompt policies for MCP tools
• Human-in-the-Loop: Configurable approval system for sensitive operations
• Audit Trail: Comprehensive logging of all command executions

See Security Model and Sandbox Deep Dive for details.

Docs & Examples

• Installation
  • Native Installers
  • Quick Reference
• Getting started
  • Interactive mode
  • Command line interface
  • Custom prompts
• Subagents
• Configuration
• AI Provider Setup - Complete guides for configuring different LLM providers:
  • OpenAI, Anthropic, Google Gemini
  • OpenRouter
  • Ollama
  • LM Studio
• Context Engineering
  • Token budget management
  • Dynamic context curation
• Agent Skills - Comprehensive skills location system with multi-location support, precedence handling, and the open Agent Skills standard.
  • Standard Compliance - How VT Code implements the Agent Skills standard
• Code Intelligence
• Agent Client Protocol (ACP)
• Zed Integration - Agent Client Protocol Integration. VT Code is fully capable ACP agent, works with ACP Clients, for example Zed.
• Lifecycle Hooks - Execute shell commands in response to agent events, enabling context enrichment, policy enforcement, and automation
• Keyboard Shortcuts
• macOS Alt Shortcut Troubleshooting - Solutions for when Alt/Option shortcuts don't work on macOS
• Custom Prompts
• Exec Mode
• Development
  • Testing
  • CI/CD
  • Empirical Evaluation - Formal framework for measuring LLM performance across safety, logic, and coding benchmarks.
• FAQ - Frequently asked questions about VT Code
• TUI & Async Guides
  • Event Handling - Terminal event architecture and patterns
  • Async Architecture - Tokio async/await design and best practices
  • Rendering Best Practices - Widget rendering and display optimization
• Ratatui Integration - Complete integration of Ratatui FAQ best practices
  • Summary - Overview of all improvements
  • FAQ Integration - Mapping of FAQ topics to code
  • Async Improvements - Tokio best practices applied
• Architecture
• Security
  • Security Model
  • Security Audit
  • Tool Policies
• Anthropic API Compatibility - Documentation for Anthropic Messages API compatibility server

Documents

• Ask Google Gemini CodeWiki
• Ask Devin DeepWiki

Visual Studio Code Extension

VT Code is available as an VS Code extension.

The original VT Code extension for Visual Studio Code with full semantic code understanding and AI assistance.

VT Code is also compatible with other VS Code-compatible editors:

Compatible with Cursor, Windsurf, and other VS Code-compatible editors through the Open VSX registry.

For installation instructions and download links for other IDEs, visit our IDE Downloads page. For troubleshooting, see the IDE Integration Troubleshooting Guide.

License

This repository is licensed under the MIT License.