webcodex

WebCodex is an open-source tool that bridges cloud-based AI agents (such as ChatGPT and Claude) with local development environments. Instead of requiring developers to upload their code to a hosted workspace, WebCodex allows AI agents to directly interact with repositories, run commands, use Git, execute tests, and perform long-running tasks—all on the developer's own machine. The tool operates through a Server/Runner architecture. The Server acts as the intermediary that AI clients connect to via MCP (Model Context Protocol) or HTTPS, while the Runner executes operations on the local machine. This design keeps the repository and all developer tools on the original machine, preserving the existing development environment. WebCodex supports two primary usage modes. For everyday development, users set up a regular Server and Runner, which provides durable access to multiple projects with full capabilities including project exploration, editing, Git operations, command execution, tests, and code navigation. For quick trials, a one-command `share` mode starts a temporary, single-project instance that prints connection values for ChatGPT. The tool supports multiple platforms: Linux (x64/arm64), macOS (x64/arm64), and Windows (x64 and arm64). Windows and macOS users are recommended to use the WebCodex Desktop application with the official OpenAI Secure Tunnel for the simplest setup. CLI-based setup is available for all platforms, with additional options for self-hosting and advanced configurations. Key capabilities include: reading, searching, inspecting, and making guarded changes to code; running commands, tests, formatters, compilers, and project-specific tooling; working with Git to inspect status and diffs; handling long-running work with observable jobs; and supporting human review through a Runtime Console, Workflow Session evidence, and Git/diff review. WebCodex is built with Rust (using Cargo for building) and is licensed under the Apache License 2.0. The project includes comprehensive documentation covering desktop installation, full setup, quick trials, MCP integration, deployment, troubleshooting, CLI reference, AI-assisted setup, and security guidance.