# Autonomous AI Agency

> A self-hosted, OpenAI-compatible AI proxy and multi-agent platform. Paste a website URL to get an audit; a CEO agent and a fleet of specialist agents then plan, write code, test it and open pull requests for human approval. MIT licensed, runs on your own servers and models.

## Links

- [Home](https://autonomous-ai-agency.strikersam.workers.dev/): what the agency does, how it works, and the proof
- [Live dashboard](https://autonomous-ai-agency.strikersam.workers.dev/login): sign in to the hosted instance
- [Source code](https://github.com/strikersam/autonomous-ai-agency): README, setup and architecture
- [Changelog](https://github.com/strikersam/autonomous-ai-agency/blob/master/CHANGELOG.md)
- [Releases](https://github.com/strikersam/autonomous-ai-agency/releases)

## What it does

- OpenAI-compatible `/v1/*`, Anthropic `/v1/messages` and Ollama `/api/*` endpoints with Bearer-token auth, rate limiting and model routing
- Plan, execute, verify agent loop driving a CEO-coordinated fleet of specialist agents
- Provider failover across NVIDIA NIM, Cerebras, Groq and Ollama
- React admin dashboard for companies, specialists, tasks, governance and system health

## FAQ

- Does it merge or deploy on its own? No. Every change is a pull request that waits for human approval; deploys, deletes, payments and messages always ask first.
- Is it free? Yes. MIT licensed and free to self-host (Python 3.11+, Node 20+, SQLite).
- Which models? Free NVIDIA NIM first, with failover across Cerebras, Groq and local Ollama.
- How to run it? `git clone https://github.com/strikersam/autonomous-ai-agency.git && cd autonomous-ai-agency && ./scripts/quickstart.sh`