Refact AI

Refact.ai is an autonomous AI coding agent, made by Small Magellanic Cloud AI Ltd., that plans and executes coding tasks end-to-end inside your IDE and codebase — in-IDE chat, real-time autocompletion, and an agent that can carry out multi-step changes across 25+ programming languages. Who it's for: developers who want an AI coding assistant they can self-host for data privacy, and teams that want to choose their own underlying LLM (Claude, GPT-4o, Gemini, and others) rather than being locked into one model provider. Strengths: on-premise deployment is a real, meaningful option for teams that can't send code to a third-party cloud, and LLM flexibility avoids vendor lock-in on the model itself. Integration with GitHub, databases, Docker, and CI/CD pipelines means it can act on a real development workflow, not just suggest snippets in isolation. Weaknesses and limitations: the company's own cloud offering (Refact Cloud) is being discontinued, with self-hosted and enterprise deployment now the emphasized path — worth knowing before planning around a hosted version. The free tier caps daily autonomous-agent usage; sustained real use needs the $10+/month Pro tier or higher. Real-world scenario: a team wants an AI coding agent but needs it self-hosted for compliance reasons, with the freedom to swap in whichever underlying LLM fits their budget and quality needs.