v0.1.0 · Rust-Powered · Open Source

OpenFang

The production-grade Agent Operating System. Not a chatbot framework — a complete OS for AI agents that runs 24/7, autonomously, while you sleep.

Get Started GitHub Repo
OpenFang Agent OS Interface - Dashboard showing autonomous agent workflows and Hands management
System Preview
180ms
Cold Start
40MB
Idle Memory
40
Channel Adapters
16
Security Layers

Seven Hands. Infinite Automation.

Each Hand is a self-contained automation unit with its own model, tools, and workflow. Think of them as AI employees that clock in at midnight and never ask for a raise.

Clip

Auto-converts YouTube videos into vertical shorts with AI captions and voice-over. Publishes directly to Telegram. Your content pipeline runs itself — no editors needed.

Lead

Daily autonomous prospect discovery with ICP-based scoring, deduplication, and CSV export. Wake up to a fresh list of qualified leads every morning.

Collector

OSINT-grade intelligence monitoring with change detection. Tracks competitors, news, and signals — alerts you only when something meaningful shifts.

Predictor

Superforecasting engine with calibrated probabilistic reasoning. Model market movements and business outcomes before your competitors do.

Researcher

Deep autonomous research with credibility evaluation. Produces structured reports with source quality scoring — not just web search summaries.

Twitter + Browser

Autonomous X/Twitter management with human approval gates. Full browser automation with purchase guardrails — power without losing control.

OpenFang vs OpenClaw vs ZeroClaw

Built in Rust. Engineered for production, not prototypes.

Cold Start

OpenClaw / ZeroClaw
2.5s – 4s+
OpenFang
180ms

Security Layers

OpenClaw / ZeroClaw
2–3 layers
OpenFang
16 layers

Channel Adapters

OpenClaw / ZeroClaw
8–15
OpenFang
40 adapters

Replace Entire Workflows with a Single Hand

OpenFang doesn't just assist — it takes over. These are the operations you'd otherwise hire people for.

video_library

Content Pipeline

Clip + Twitter working together: monitor trending videos, cut shorts, add captions, publish to social — all while you're offline.

filter_alt

Sales Prospecting

Lead runs nightly: discovers prospects, scores by ICP fit, removes duplicates, exports a clean CSV. Wake up ready to sell.

monitoring

Competitive Intelligence

Collector watches competitor sites, pricing, job boards, and news. Alerts you the moment something changes — not after you notice it.

hub

Multi-Agent Orchestration

Chain Hands with workflow orchestration. Build pipelines like: Researcher → Predictor → Clip → broadcast to 40 channels in one run.

sync_alt

Migrate from OpenClaw

One command migration: openfang migrate --from openclaw. Your agents, memory, and skills transfer without manual reconfiguration.

security

Production Security

WASM sandbox, Merkle audit chain, SSRF protection, prompt injection scanning, GCRA rate limiting — 16 layers so you can deploy with confidence.

One Command. Then Walk Away.

Single binary. No Docker required. Runs on your laptop, a VPS, or a Raspberry Pi. Start autonomous agents in under 60 seconds.

bash — Quick Install
curl -fsSL https://openfang.sh/install | sh
checklist

Requirements

  • • Linux / macOS / Windows
  • • 64MB RAM minimum (256MB recommended)
  • • x86_64 or ARM64 architecture
  • • LLM API Key (12 providers supported)
settings_suggest

What You Get

  • • 7 built-in autonomous Hands
  • • 10 workflow orchestration templates
  • • 40 channel adapters (incl. Feishu & DingTalk)
  • • Tauri 2.0 native desktop app

Three Steps to Autonomous AI

  1. 1

    Install

    curl -fsSL https://openfang.sh/install | sh
# Windows PowerShell:
irm https://openfang.sh/install.ps1 | iex
  2. 2

    Initialize

    openfang init
# Configure your LLM provider and channel tokens

    Follow the interactive setup to connect your LLM API key and messaging channels

  3. 3

    Start Agents

    openfang start
# Migrate from OpenClaw:
openfang migrate --from openclaw

    Agents run as background daemons. Use the desktop app or web UI to monitor and manage.

macOS
Linux
Windows
Raspberry Pi
VPS / Cloud

Frequently Asked Questions

What is OpenFang?
OpenFang is a production-grade Agent Operating System built in Rust. Unlike chatbot frameworks, it's designed to run autonomous agents 24/7 on schedules — without requiring user prompts. Think of it as the OS layer beneath your AI agents, providing runtime, security, and channel infrastructure.

How is it different from OpenClaw?
OpenClaw is great for personal use and experimentation. OpenFang is what you reach for when you need production-grade reliability — Rust instead of TypeScript, 16 security layers vs basic sandboxing, built-in Hands instead of plugin glue code. OpenFang also includes a built-in migration path: openfang migrate --from openclaw.

What are Hands?
Hands are OpenFang's autonomous capability units — think of them as enhanced plugins, but self-contained. Each Hand ships with a designated model, tools, and workflow. You don't wire anything together; you just activate the Hand. The seven built-in Hands cover video (Clip), prospecting (Lead), intelligence (Collector), forecasting (Predictor), research (Researcher), social media (Twitter), and web automation (Browser).

Which LLM providers are supported?
OpenFang supports 12 mainstream LLM providers including Anthropic, OpenAI, Gemini, Groq, DeepSeek, Mistral, Together, Ollama, vLLM, and more — covering 123+ models total. All providers require API keys; OAuth is not currently supported. Each Hand can be configured to use a different provider.

Which channels are supported?
OpenFang supports 40 channel adapters — the broadest coverage of any agent framework. This includes Telegram, Discord, Slack, WhatsApp, Signal, Matrix, Teams, Google Chat, Feishu, DingTalk, Mastodon, Bluesky, LinkedIn, Reddit, IRC, and 24+ more. If your team uses it, OpenFang probably connects to it.

Is it production-ready?
OpenFang v0.1.0 passes 1,767+ tests with zero Clippy warnings. It has 16 discrete security layers including WASM sandboxing, Merkle hash-chain audit trail, and SSRF protection. Breaking API changes are possible until v1.0, so pin your version in production. The core runtime is stable and used in production deployments today.

Where is the GitHub repository?
Find OpenFang on GitHub: github.com/RightNow-AI/openfang. Star the project to follow releases and feature updates.