Skip to main content

System requirements

OpenFang is a single ~32MB binary with zero runtime dependencies. It runs on:
  • Operating Systems: macOS (10.15+), Linux (glibc 2.31+), Windows (10/11), Docker
  • Architectures: x86_64 (Intel/AMD), ARM64 (Apple Silicon, AWS Graviton)
  • Memory: 100MB minimum, 512MB recommended
  • Disk: 100MB for binary + data
  • Network: Optional (for LLM API calls and channel integrations)
OpenFang can run fully offline if you use a local LLM provider like Ollama.

Installation methods

Choose the method that works best for your platform:

One-line installer

Recommended for macOS/Linux

PowerShell installer

Recommended for Windows

Docker

For containerized deployments

Build from source

For contributors and custom builds

macOS installation

The installer downloads the latest pre-built binary, verifies the checksum, and adds it to your PATH.
What it does:
  1. Detects your architecture (Intel x86_64 or Apple Silicon ARM64)
  2. Downloads the latest release from GitHub
  3. Verifies SHA256 checksum
  4. Installs to ~/.openfang/bin/openfang
  5. Adds ~/.openfang/bin to your PATH (in ~/.zshrc or ~/.bashrc)
Apple Silicon users: The installer automatically downloads the ARM64 binary optimized for M1/M2/M3 chips.

Homebrew (coming soon)

Homebrew formula is in progress. Use the one-line installer for now.

Verify installation

Expected output:

Update OpenFang

Re-run the installer to upgrade to the latest version:
Or manually download a specific version:

Linux installation

Works on Ubuntu, Debian, Fedora, Arch, and most glibc-based distributions:
Supported distributions:
  • Ubuntu 20.04+ (Focal, Jammy, Noble)
  • Debian 11+ (Bullseye, Bookworm)
  • Fedora 36+
  • Arch Linux (rolling)
  • RHEL/CentOS 8+
  • Amazon Linux 2023
  • Any glibc 2.31+ distribution

Install to a custom directory

If you install to a system directory like /usr/local/bin, you may need sudo privileges.

Package managers

Official packages are in progress. Use the one-line installer for now.

Verify installation

Add to systemd (run as a service)

Create a systemd service file to run OpenFang automatically on boot:
Enable and start the service:
Check status:

Windows installation

Run this command in PowerShell (right-click → Run as Administrator if needed):
What it does:
  1. Detects your architecture (x64 or ARM64)
  2. Downloads the latest release (.zip)
  3. Verifies SHA256 checksum
  4. Extracts openfang.exe to %USERPROFILE%\.openfang\bin
  5. Adds the directory to your user PATH
Restart your terminal (CMD, PowerShell, or Git Bash) after installation for PATH changes to take effect.

Windows Subsystem for Linux (WSL)

If you’re using WSL2, follow the Linux installation instructions inside your WSL environment.

Verify installation

Open a new PowerShell or CMD window and run:
Expected output:

MSI installer (coming soon)

Double-click to install with a GUI wizard:
MSI installer is in development. Use the PowerShell script for now.

Run as a Windows service

Use NSSM (Non-Sucking Service Manager) to run OpenFang as a background service:

Docker installation

Official image

Pull the latest OpenFang image from Docker Hub:

Run OpenFang in a container

Start the daemon with persistent storage:
The -v openfang-data:/data volume persists agent configurations, conversations, and memory between container restarts.

Access the dashboard

Open http://localhost:4200 in your browser.

Run a one-off command

Execute OpenFang CLI commands inside the container:

Docker Compose

Create a docker-compose.yml file:
docker-compose.yml
Start the service:

Build your own Docker image

Clone the repository and build locally:

Build from source

Prerequisites

Install Rust 1.75 or later:

Clone the repository

Build the project

OpenFang is a Cargo workspace with 14 crates:
The first build takes 5-10 minutes because it compiles SQLite (bundled) and Wasmtime. Subsequent builds are incremental and much faster.

Install the binary

Copy the compiled binary to a directory in your PATH:

Run tests

OpenFang has 1,744+ tests covering all subsystems:

Run clippy (zero warnings policy)

Development build

For faster compile times during development:

Configuration

After installation, initialize OpenFang:
This creates:
  • ~/.openfang/config.toml — Main configuration file
  • ~/.openfang/.env — Environment variables for API keys
  • ~/.openfang/data/openfang.db — SQLite database
  • ~/.openfang/logs/ — Log files

Example configuration

~/.openfang/config.toml:
~/.openfang/.env:

Environment variables

OpenFang respects the following environment variables: Set them in your shell profile or in ~/.openfang/.env for persistence.

Updating OpenFang

Update via installer

Re-run the installation script:

Update Docker image

Update from source


Uninstalling OpenFang

Remove the binary

Or use the built-in uninstaller

This removes:
  • The binary (~/.openfang/bin/openfang)
  • Configuration files (~/.openfang/config.toml)
  • Database and logs (~/.openfang/data/)
  • PATH entries from shell config files
The uninstaller permanently deletes all agent data, conversations, and memory. Back up ~/.openfang/data/ first if you want to preserve it.

Troubleshooting

The installer requires HTTPS to download from GitHub. Ensure you have up-to-date SSL certificates.Fix (Ubuntu/Debian):
Fix (macOS):
The binary may not have execute permissions.Fix:
The installer added ~/.openfang/bin to your PATH, but you may need to restart your terminal.Fix:
The download may be corrupted or tampered with.Fix:
If the issue persists, report it on GitHub.
Check the container logs for errors:
Common causes:
  • Missing API key: Add -e GROQ_API_KEY=... to the docker run command
  • Port conflict: Change -p 4200:4200 to -p 5000:4200
Ensure you have Rust 1.75+ installed:
If the build fails, check CONTRIBUTING.md for detailed build instructions.

Next steps

Now that OpenFang is installed, continue with the Quickstart to:
  • Initialize configuration with openfang init
  • Start the daemon with openfang start
  • Spawn your first agent with openfang agent spawn researcher
  • Chat with the agent using openfang chat researcher

Quickstart Guide

Get your first agent running in 5 minutes

Configuration

Configure models, channels, and security

Creating Agents

Create custom agents with templates

API Reference

Integrate OpenFang programmatically

Get help

Discord

Ask questions and get support

GitHub Issues

Report bugs or request features

Twitter/X

Follow for updates