Graphite lets coding agents ask questions about how code fits together
An open-source tool called Graphite gives AI coding agents a searchable map of a codebase, so they can trace callers and values instead of guessing.
The short version
- Graphite is open source under the Apache-2.0 license
- It connects to coding agents through MCP and exposes a program graph via Cypher
- Agents can find callers, trace values across methods and see where config keys are read
- Supports JVM artifacts (jar/war/apk), compiled JavaScript/TypeScript, and Swift/Xcode projects
Why it matters
AI coding assistants can only read so much at once, so on big projects they often guess which code matters. A tool that answers precise structural questions replaces some of that guessing with lookups. The idea also carries beyond code: assistants do better when they can query organized information rather than read everything.
What Graphite does
As dev.to reported, a developer is building Graphite, an open-source tool that lets coding agents ask questions about the structure of a software project. Instead of reading files one by one, an agent can query a map of how the pieces connect.
The author says it can find which code calls a given function, trace a value as it moves across methods, and check where configuration settings are read.
How the pieces fit
Graphite builds what the author calls a program graph: a web of code elements and the links between them. Agents query it using Cypher, a query language designed for graph data.
The connection to the agent runs through MCP, short for Model Context Protocol, a common way for AI tools to plug into outside programs and data. Any agent that speaks MCP can, in principle, use it.
What it can read
The post lists several kinds of projects Graphite supports:
- JVM artifacts, meaning jar, war and apk files
- Compiled JavaScript and TypeScript
- Swift and Xcode projects
It is released under the Apache-2.0 license, which is permissive and allows broad reuse. The code is on GitHub under johnsonlee/graphite.
What we do not know yet
The post is short. It does not say how well Graphite performs on large projects, how much it improves an agent's answers, or how hard it is to set up. Treat it as an early project announcement by its creator, not a tested result.
What to watch
- Whether independent tests show agents give more accurate answers with Graphite
- Whether support grows beyond the JVM, JavaScript/TypeScript and Swift/Xcode projects listed
Try it
If you code with an AI assistant, notice where it guesses about callers or settings. Those are the questions a tool like Graphite targets; check the GitHub page to see if your stack is supported.
Sources and links
- Reportingdev.to: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by dev.to · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
