What is codegraph?

CodeGraph is a tool that builds a pre-indexed map of your codebase and makes that map available to Claude Code, the AI coding assistant. The idea is to solve a slow and expensive problem: when Claude Code needs to understand a project, it normally explores files one by one using search and read operations, each of which costs tokens (the units AI services charge for). CodeGraph does that exploration work upfront, offline, and stores the results in a local database, so when Claude Code needs to understand your code it queries the pre-built map instead of reading files on demand. The practical result, according to benchmarks shown in the README, is roughly 90 percent fewer tool calls and 70 percent faster responses when Claude Code answers questions about a codebase. The graph tracks which functions call which other functions, where symbols are defined, and how code is connected across files. It also understands routing patterns in web frameworks, so it can link a URL path back to the function that handles it. Everything runs on your own machine, no data is sent anywhere, no API keys are required, and the database is just a local SQLite file (a lightweight database that lives as a single file on disk). A file watcher keeps the graph up to date as you edit code, so you do not need to re-index manually. You set it up by running one command, which installs the tool, configures Claude Code to use it automatically, and indexes your project. It supports over 19 programming languages. The project is built with TypeScript and runs on Node.js. The full README is longer than what was provided.