archestra

Secure MCP runtime for fully autonomous agents as a desktop application

GitHub Stars

96

User Rating

Not Rated

Favorites

0

Views

13

Forks

15

Issues

15

README

⚠️ Early Development Notice

Archestra is in very early development, and it's not yet ready for general download.

Want to get notified when we launch? Drop your email at Archestra.AI to be the first to know when Archestra is ready.

🚀 Overview

Archestra is an enterprise-grade platform that enables non-technical users to safely leverage AI agents and MCP (Model Context Protocol) servers. Built as a privacy-focused desktop application, Archestra provides a secure runtime environment for AI interactions with sandboxing, resource controls, and prompt injection prevention.

Archestra Screenshot

Key Features
  • 🔒 Secure Runtime - Isolated execution environment for AI agents with sandboxing and resource controls
  • 🛡️ Prompt Injection Prevention - Advanced security to prevent context leakage and behavior manipulation
  • 🔌 MCP Protocol Support - Compatible with thousands of existing MCP servers and custom implementations
  • 💻 Local-First Architecture - Privacy-focused design with local LLM
  • 🎯 Enterprise Ready - Built for non-technical users with enterprise-level security and compliance
  • 🔧 Extensible Tool System - Support most of MCP servers: GitHub, Gmail, Slack, PostgreSQL, filesystem, and more
🤝 Contributing

We welcome contributions from the community! Contribution guideline.



Linux Foundation        CNCF