explaingit

rightnow-ai/openfang

17,495RustAudience · developerComplexity · 3/5LicenseSetup · easy

TLDR

OpenFang is an open-source Agent Operating System written in Rust that runs autonomous AI agents on a schedule in the background, covering tasks like video clipping, lead finding, web automation, and research reports via a local dashboard.

Mindmap

mindmap
  root((OpenFang))
    What it is
      Agent OS
      Rust single binary
      Local dashboard
    Hands
      Clip video clips
      Lead sales leads
      Researcher reports
      Browser web tasks
      Twitter X account
      Predictor forecasts
      Collector knowledge
    Features
      Scheduled runs
      Approval gates
      HAND.toml config
    Audience
      Developers
      Automation users
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Run an overnight agent that finds and scores sales leads each day and delivers a report to your local dashboard without manual effort.

USE CASE 2

Automatically clip a YouTube video into short vertical social media clips with captions using the built-in Clip Hand.

USE CASE 3

Set up a research agent that monitors a topic, builds a knowledge graph, and generates cited reports on a recurring schedule.

Tech stack

Rust

Getting it running

Difficulty · easy Time to first run · 30min

Installs as a single 32MB Rust binary, individual Hands may require API keys for the LLM or external services they use.

MIT license, use freely for any purpose, including commercial use, as long as you keep the copyright notice.

In plain English

OpenFang is described as an open-source Agent Operating System: software for running autonomous AI agents that do work for you in the background instead of waiting for you to type a prompt. It is written in Rust and ships as a single binary of roughly 32MB, so installing and starting it is a one-line command that brings up a local dashboard. The README contrasts this with chatbot frameworks and multi-agent orchestrators that require heavier setups. The core idea is what the README calls Hands. A Hand is a pre-built capability package that bundles a configuration file (HAND.toml), a multi-step system prompt, a domain skill reference, and guardrails for approval gates. Once activated, a Hand runs on a schedule and reports results back to the dashboard. Seven Hands are bundled: Clip, which turns YouTube videos into vertical short clips with captions, Lead, which finds and scores sales prospects each day, Collector, which continuously monitors a target and builds a knowledge graph, Predictor, which makes forecasts with confidence intervals and tracks its own accuracy, Researcher, which produces cited reports across multiple sources, Twitter, which manages an X account with an approval queue, and Browser, which automates web tasks via a Playwright bridge and refuses to spend money without explicit confirmation. Someone would use OpenFang if they want agents that run on their own (for example overnight) rather than a chat interface. The stated stack is Rust, MIT-licensed, and the project is still pre-1.0 with breaking changes possible between minor versions. The full README is longer than what was provided.

Copy-paste prompts

Prompt 1
I have installed OpenFang. Walk me through activating the Lead Hand to search for SaaS prospects daily and send results to my dashboard.
Prompt 2
How do I configure the Clip Hand in OpenFang to take a YouTube URL and output vertical clips under 60 seconds with auto-generated captions?
Prompt 3
I want OpenFang's Researcher Hand to monitor a competitor and produce a weekly summary report. Write the HAND.toml config for this task.
Prompt 4
Explain how OpenFang's approval gates work so I can confirm the Browser Hand will never spend money or submit forms without my explicit confirmation.
Open on GitHub → Explain another repo

← rightnow-ai on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.