Open Interpreter Returns as OpenAI Codex Fork With 64,900 GitHub Stars and Major New Features
Summary
Open Interpreter is back as a fork of OpenAI's Codex, now boasting 64,900 GitHub stars, a lightweight Rust-based architecture, computer use capabilities, and support for top open models like DeepSeek and Kimi across macOS, Linux, and Windows.
Key Points
- Open Interpreter, a lightweight Rust-based coding agent optimized for low-cost open models like GLM, Deepseek, and Kimi, is now back with a major update as a fork of OpenAI's Codex, boasting over 64,900 GitHub stars and 512 contributors.
- The tool supports multiple agent harnesses including native, claude-code, swe-agent, and others, and ships with computer use capabilities that allow any model to operate web and native app interfaces through built-in QA skills.
- Available on macOS, Linux, and Windows via simple install commands, Open Interpreter runs commands in native sandboxes, supports MCP, hooks, skills, and AGENTS.md, and keeps all config and session state stored locally under ~/.openinterpreter.