OpenFang
A self-hosted AI agent platform for macOS, Windows and Linux, with scheduled automation, Ollama support and an Apache 2.0 license.
OpenFang runs scheduled AI agents on your own machine, with built-in task packages called Hands that report results to a dashboard. It's for people who want recurring research, monitoring or business tasks to continue without prompting an agent each time. The open source platform runs on macOS, Windows and Linux under the Apache 2.0 license.
Hands cover concrete jobs. Researcher checks sources and produces cited reports; Collector monitors targets for changes and builds knowledge graphs. Lead finds and scores prospects, while Clip turns longer videos into captioned shorts. Other Hands handle forecasting, X account management and browser workflows. The browser agent requires approval before making purchases, and you can create custom Hands for other tasks.
The Rust runtime ships as a single binary with a native desktop dashboard. SQLite-backed memory retains context across conversations and messaging channels, including Telegram, Discord, Slack and WhatsApp. MCP support connects external tools and makes OpenFang's tools available to other agents; its workflow engine supports multi-agent pipelines with conditional steps and loops.
Model connections include Ollama, vLLM and LM Studio for local inference, alongside cloud providers such as Anthropic, Gemini and OpenAI. The runtime and memory run locally; choosing a cloud model sends model requests to that provider. It includes a WebAssembly sandbox, workspace limits on file access, prompt injection scanning and a cryptographically linked audit trail. OpenFang is still a pre-1.0 project.
