MyAgents is an open-source desktop AI Agent application designed to democratize AI capabilities for a wide range of users, including students, creators, and professionals. It provides a unified workbench and task center, acting as a personal AI assistant to amplify user intent and productivity, with a strong emphasis on local data privacy and ease of use.

How It Works

The application employs a session-centric, multi-instance Sidecar architecture. Each user tab runs an isolated Agent process, managed by a Rust proxy layer that handles all inter-process communication and eliminates CORS issues. A multi-owner mechanism allows Tabs, scheduled tasks, and external channels to safely share underlying Agent Sidecars. It utilizes Tauri v2 (Rust) for the desktop framework and React/TypeScript for the frontend. The Agent runtime is powered by Bun, defaulting to the Claude Agent SDK but extensible to Claude Code CLI or OpenAI Codex CLI, while Node.js supports the MCP server and community npm ecosystem.

Quick Start & Requirements

• Installation: Download the installer directly from https://myagents.io.
• System Requirements:
  • End Users: macOS 13.0 (Ventura) or later, Windows 10 or later.
  • Developers: macOS 13.0+ / Windows 10+, Node.js (v18+), Bun, Rust.
• Relevant Links: Official Website & Downloads: https://myagents.io. Contribution Guidelines: CONTRIBUTING.md.

Highlighted Details

• Features a Chrome-style multi-tab GUI for parallel Agent workflows.
• Supports over 10 AI model providers, including Anthropic, DeepSeek, Moonshot, Zhipu AI, MiniMax, and OpenRouter, allowing cost-controlled model selection.
• Includes a Skills System for reusable Agent capabilities and MCP Tool Integration for unlimited extensibility.
• Offers an Agent + Channel architecture with built-in Telegram/DingTalk adapters and plugin support for other IM platforms.
• Integrates an embedded terminal (PTY) and a preview browser within the GUI.
• Provides local full-text search over session history and workspace files using Tantivy and jieba.
• All user data (conversations, files, memories) is stored locally, ensuring privacy.

Maintenance & Community

The project provides contribution guidelines via CONTRIBUTING.md. Specific details regarding active maintainers, community channels (like Discord/Slack), or a public roadmap are not explicitly detailed in the README.

Licensing & Compatibility

The project is licensed under the Apache License 2.0. This license is permissive and generally compatible with commercial use and linking within closed-source projects.

Limitations & Caveats

The "Multi-Agent Runtime (Lab)" designation suggests some runtime configurations may be experimental. The rapid evolution of AI technology implies potential for frequent updates and feature changes. No specific alpha status, known bugs, or deprecations are highlighted in the provided README.