OpenFang

OpenFang is an open-source Agent Operating System (Agent OS) built entirely in Rust, designed to run, manage, and deploy autonomous AI agents at production scale. It ships as a single binary containing 14 crates, 137K lines of Rust, 30 pre-built agents, 40 channel adapters, 38 native tools, and 16 discrete security systems. Unlike traditional agent frameworks, OpenFang introduces "Hands" — pre-built autonomous capability packages that run on schedules, build knowledge graphs, and report results to a dashboard without requiring user prompts. It supports 27 LLM providers and integrates Model Context Protocol (MCP), Google A2A, and the OpenFang Protocol (OFP) for P2P agent networking.

• 7 Autonomous Hands — Pre-built capability packages (Clip, Lead, Collector, Predictor, Researcher, Twitter, Browser) that work autonomously on schedules and report to your dashboard.
• 16 Security Systems — WASM dual-metered sandbox, Ed25519 manifest signing, Merkle audit trail, taint tracking, SSRF protection, HMAC-SHA256 mutual auth, prompt injection scanner, and more.
• 40 Channel Adapters — Connect agents to Telegram, Discord, Slack, WhatsApp, Teams, IRC, Matrix, and 33 additional platforms with per-channel model overrides and rate limiting.
• 30 Pre-Built Agents — Orchestrators, code reviewers, customer support agents, and more across four performance tiers spanning Anthropic, Gemini, Groq, and DeepSeek.
• 38 Built-In Tools + MCP — Native tools including web search, browser automation, image generation, TTS, Docker, and knowledge graphs, plus full MCP client and server support.
• Persistent Memory — SQLite-backed storage with vector embeddings, cross-channel canonical sessions, automatic LLM-based compaction, and JSONL session mirroring.
• Tauri 2.0 Desktop App — Native desktop application with system tray, notifications, single-instance enforcement, auto-start on login, and global shortcuts.
• Multi-Agent Workflows — Fan-out, conditional, and loop step modes for building complex multi-agent pipelines with a built-in workflow engine.
• MCP + A2A + OFP Protocols — Full Model Context Protocol client/server, Google Agent-to-Agent tasks, and OpenFang Protocol for authenticated P2P agent networking.
• Quick Install — Install via a single curl command (curl -fsSL https://openfang.sh/install | sh) on macOS, Linux, or Windows, then spawn your first agent in under 2 minutes.