OpenHands

OpenHands, from All Hands AI, is the leading open-source AI coding agent. It edits code, runs commands, browses the web, and completes development tasks inside a sandboxed environment — and it consistently scores near the top of SWE-bench among open agents, which is the benchmark most people watch for real-world coding ability. Crucially, it is MIT-licensed, so you can read, modify, and run the whole thing yourself.

The economic model is what draws many users: there is no subscription. You self-host OpenHands and bring your own model API key, so you pay only for the model calls you actually make. That gives you full control, auditability, and the ability to customize the agent's behavior — valuable for privacy-sensitive teams that cannot send code to a third-party SaaS, and for researchers who want to experiment with the agent loop itself.

The trade-off is polish and setup. Running OpenHands well requires technical comfort: you handle the deployment, the sandbox, and the model configuration yourself, and the experience is less buttoned-up than commercial rivals. There is a hosted cloud option if you would rather not self-host. For budget-conscious developers, privacy-sensitive teams, and tinkerers, the combination of zero subscription cost and complete control is hard to beat.