wtf is fastctx?

FastCtx is a local tool runtime, written in Rust, that gives AI coding agents like ChatGPT or Codex a faster and cleaner way to work with a code repository. Instead of having the AI construct raw shell commands to read files, search text, or list files, which forces it to worry about quoting, escaping, and platform differences, FastCtx exposes those repository operations as structured tools that take simple parameters such as a path, a pattern, or a range. The problem it addresses is that coding agents spend a lot of their limited attention on tool mechanics rather than the actual code. Building and checking a shell command correctly can take several back and forth tool calls before useful information comes back, and that overhead eats into the context and reasoning the model has available for the real task. FastCtx moves that work into a persistent Rust process with stable input and output formats, so the model can gather what it needs in fewer steps. It ships nine tools through the Model Context Protocol: read for text, images, PDFs, and raw bytes, grep for searching file contents, glob for finding files by pattern, replace for mechanical batch text replacement, and a set of run tools for executing Bash commands in the foreground or as tracked background jobs that can be listed, followed, or killed. Installation is through npm, requiring Node.js 18 or later, and running the fastctx command opens a full screen control terminal supporting seventeen languages where you review and apply configuration changes before restarting your ChatGPT or Codex session. The tool can also be installed with cargo or downloaded directly from GitHub Releases for Windows, Linux, and macOS. FastCtx checks for updates in the background, verifies downloaded release archives against published checksums, and can roll back an update automatically if a restart fails. A command line mode supports non interactive setup, status checks, and background job management for scripting.