easi.run gives AI agents real Chrome on your own Linux server. When a task stops at login, 2FA or an approval, watch live, take control for the human step, then hand the browser back so the agent can continue and download the result.

 

For developers and technical operators using MCP clients or coding agents. Connect your own agent through MCP or REST, reuse encrypted saved browser profiles, upload input files and receive downloads on your device. Record sessions and review the action timeline.

 

The software is open source under AGPL-3.0, with no per-session usage fee. You provide Linux x86-64, Docker Engine and your AI model; server and model costs remain.

 

Code: https://github.com/julianosig/easi.run

Docs: https://docs.easi.run

Demo: https://easi.run/demo

Official video: https://easi.run/film/en-997a5b9681d22834.mp4

Maker: Juliano Gouveia, Shimpz. Contact: julianoamg.dev@gmail.com