Skip to main content

​
The Agent Operating System

Open-source Agent OS built in Rust. 137K LOC. 14 crates. 1,767+ tests. Zero clippy warnings. One binary. Battle-tested. Agents that actually work for you.

​
What is OpenFang?

OpenFang is an open-source Agent Operating System — not a chatbot framework, not a Python wrapper around an LLM, not a “multi-agent orchestrator.” It is a full operating system for autonomous agents, built from scratch in Rust. Traditional agent frameworks wait for you to type something. OpenFang runs autonomous agents that work for you — on schedules, 24/7, building knowledge graphs, monitoring targets, generating leads, managing your social media, and reporting results to your dashboard. The entire system compiles to a single ~32MB binary. One install, one command, your agents are live.

7 Autonomous Hands

Pre-built capability packages that run independently on schedules

40 Channel Adapters

Telegram, Discord, Slack, WhatsApp, Matrix, Signal, and 34 more

27 LLM Providers

123+ models from Anthropic, OpenAI, Gemini, Groq, DeepSeek, and more

16 Security Layers

WASM sandbox, Merkle audit trail, HMAC auth, capability gates

53 Built-in Tools

Plus MCP and A2A protocol support for external integrations

Single Binary

32MB with zero dependencies - one install, one command

​
Quick Start

1

Install OpenFang

curl -fsSL https://openfang.sh/install | sh
2

Initialize and Configure

Set up your configuration and API keys:
openfang init

# Set your LLM provider API key
export ANTHROPIC_API_KEY=sk-ant-...
# or OPENAI_API_KEY, GROQ_API_KEY, etc.
3

Start the Daemon

Launch the OpenFang daemon:
openfang start
# Dashboard live at http://localhost:4200
4

Activate a Hand

Start an autonomous agent working for you:
# Activate the Researcher Hand
openfang hand activate researcher

# Or spawn a custom agent
openfang agent spawn coder

​
Core Features

Autonomous Hands

7 pre-built autonomous capability packages: Researcher, Lead Gen, Collector, Predictor, Twitter, Browser, and Clip

Architecture

14 Rust crates, modular kernel design, SQLite memory substrate with vector embeddings

Security

16 discrete security systems including WASM sandbox, capability gates, and Merkle audit trail

Channel Adapters

Connect to 40 messaging platforms with DM/group policies and rate limiting

LLM Providers

27 providers with 123+ models, intelligent routing, and cost tracking

API Reference

140+ REST/WebSocket endpoints including OpenAI-compatible interface

​
Hands: Agents That Actually Do Things

Hands are OpenFang’s core innovation — pre-built autonomous capability packages that run independently, on schedules, without you having to prompt them.

Researcher Hand

Deep autonomous researcher with cross-referencing, fact-checking, and structured reports

Lead Hand

Daily lead generation that discovers prospects, enriches data, and scores 0-100

Collector Hand

OSINT-grade intelligence with continuous monitoring and knowledge graph construction

Predictor Hand

Superforecasting engine with calibrated reasoning and Brier score tracking

Twitter Hand

Autonomous Twitter/X manager with 7 content formats and approval queues

Browser Hand

Web automation with Playwright bridge and mandatory purchase approval gates

Clip Hand

YouTube to vertical shorts pipeline with captions, thumbnails, and AI voice-over

Build Your Own

Define HAND.toml with tools, settings, and system prompt. Publish to FangHub.

​
Why OpenFang?

  • Cold Start: <200ms (vs. 3-6s for Python frameworks)
  • Memory: 40MB idle (vs. 180-400MB for others)
  • Install Size: 32MB binary (vs. 100-500MB)
  • Tests: 1,767+ passing with zero clippy warnings
  • WASM dual-metered sandbox (fuel + epoch interruption)
  • Merkle hash-chain audit trail (tamper-evident logging)
  • Ed25519 signed agent manifests
  • HMAC-SHA256 mutual authentication for P2P
  • SSRF protection and secret zeroization
  • Capability-based access control with inheritance validation
  • 40 channel adapters across all major messaging platforms
  • 60 bundled skills (GitHub, Docker, K8s, AWS, and more)
  • MCP and A2A protocol support for external integrations
  • Desktop app (Tauri 2.0) with system tray and notifications
  • OpenAI-compatible API for drop-in replacement
  • Single binary with zero dependencies
  • Comprehensive CLI with daemon auto-detect
  • Hot-reloadable configuration
  • Built-in migration from OpenClaw
  • Extensive documentation and 30 agent templates

​
Next Steps

Quickstart Guide

Get up and running in 5 minutes

Installation

Detailed installation instructions for all platforms

Core Concepts

Understand the architecture and design

CLI Reference

Complete command-line interface documentation

Build docs developers (and LLMs) love

Get started for freeTalk to us