I built Donovan Agent because I wanted a more practical way to work with AI locally, not just as a chatbot, but as a tool that could actually help with files, code, tasks, and structured workflows. The problem I was trying to solve was that most AI tools feel either too limited, too closed, or too disconnected from the user’s real environment. I wanted Donovan to feel like a local AI command center that gives users more control, transparency, and flexibility. The process evolved a lot while building it. At first, it was more of an experiment, but over four months it became a much more serious open-source project. Later, I added Qwen 3.6 35B into my workflow as a coding copilot, which massively improved my productivity and helped me move faster while still keeping the core design decisions and direction my own.

Donovan Agent

The self-evolving computer agent

The self-evolving computer agent