CodeGraph v2: 38K Stars Update — Token Optimization, New Platforms, and 5x Growth

Two weeks ago, we covered CodeGraph when it had 7,600 stars. Today it has 38,745 — a 5x explosion driven by one insight: fewer tokens, fewer tool calls, 100% local. This update covers what changed, what’s new, and why the developer community is adopting CodeGraph at breakneck speed.

The Growth Story

CodeGraph went from 7,600 to 38,745 stars in approximately two weeks, with +10,793 stars this week alone. That makes it one of the fastest-growing developer tools on GitHub.

Why the explosive growth? Three reasons:

1. Token efficiency matters more than ever. As AI coding agents proliferate, context window waste is the bottleneck. CodeGraph eliminates wasteful file-scanning loops.
2. Platform expansion. CodeGraph now supports 8+ AI coding agent platforms, up from 4 in the original release.
3. Zero configuration. One command installs and configures everything. No API keys, no external services, no data leaving your machine.

Token Optimization: The New Focus

The original CodeGraph pitch was “94% fewer tool calls, 77% faster exploration.” The v2 message sharpens the focus: fewer tokens consumed.

Why Tokens Matter

Every tool call an AI agent makes consumes tokens — both in the request and the response. When Claude Code explores an unfamiliar codebase, it spawns Explore agents that scan files using grep, glob, and Read tool calls. For a large project like VS Code, answering a single architecture question can require 52 tool calls consuming 50K+ tokens.

CodeGraph replaces this wasteful discovery loop with structured graph queries. The same VS Code architecture question takes 3 tool calls and ~2K tokens with CodeGraph.

How CodeGraph Reduces Token Consumption

Approach	Tool Calls	Tokens	Time
Without CodeGraph	52 calls	50K+ tokens	1m 37s
With CodeGraph	3 calls	~2K tokens	17s
Improvement	94% fewer	~96% fewer	82% faster

The codegraph_context tool is the token-saving powerhouse. It combines search, navigation, and code retrieval into a single call that returns everything an agent needs to understand a code area. This replaces the 40-50 tool calls that agents normally make during exploration.

Expanded Platform Ecosystem

CodeGraph v2 expands from 4 to 8+ supported AI coding agent platforms:

Platform	Status	Integration
Claude Code	Original	MCP server + CLAUDE.md instructions
Cursor	Original	MCP server + .cursor/rules/codegraph.mdc
Codex CLI	Original	MCP server + AGENTS.md instructions
opencode	Original	MCP server integration
Gemini CLI	New	MCP server integration
AntiGravity	New	MCP server integration
Kiro	New	MCP server integration
Hermes Agent	New	MCP server integration

All platforms connect through the same MCP server interface. The npx @colbymchenry/codegraph installer auto-detects installed agents and configures each one appropriately.

Architecture Overview

CodeGraph builds a semantic knowledge graph of your codebase using tree-sitter AST parsing, stores it in a local SQLite database with FTS5 full-text search, and exposes it to AI agents through an MCP server with 8 specialized tools.

The architecture follows a 5-layer pipeline:

1. ExtractionOrchestrator — tree-sitter parses source code into ASTs; language-specific queries extract 22 NodeKinds and 12 EdgeKinds
2. ReferenceResolver — connects function calls to definitions, resolves imports, establishes inheritance chains
3. Framework Detection — recognizes web-framework routing files across 13 frameworks (Django, Flask, FastAPI, Express, Laravel, Rails, Spring, Gin, chi, gorilla/mux, Axum, actix, Rocket, ASP.NET, Vapor, React Router, SvelteKit)
4. SQLite Storage — local .codegraph/codegraph.db with FTS5 full-text search, using better-sqlite3 (native) or node-sqlite3-wasm (fallback)
5. Auto-Sync — native OS file events (FSEvents/inotify/ReadDirectoryChangesW) with 2-second debounced incremental sync

Benchmark Results

Tested across 6 real-world codebases comparing Claude Code’s Explore agent with and without CodeGraph:

Codebase	With CodeGraph	Without CodeGraph	Improvement
VS Code (TypeScript)	3 calls, 17s	52 calls, 1m 37s	94% fewer, 82% faster
Excalidraw (TypeScript)	3 calls, 29s	47 calls, 1m 45s	94% fewer, 72% faster
Claude Code (Python+Rust)	3 calls, 39s	40 calls, 1m 8s	93% fewer, 43% faster
Claude Code (Java)	1 call, 19s	26 calls, 1m 22s	96% fewer, 77% faster
Alamofire (Swift)	3 calls, 22s	32 calls, 1m 39s	91% fewer, 78% faster
Swift Compiler (Swift/C++)	6 calls, 35s	37 calls, 2m 8s	84% fewer, 73% faster

Key observation: With CodeGraph, agents never fell back to reading files — they trusted the graph results completely. The Swift Compiler benchmark tested the largest codebase (25,874 files, 272,898 nodes) and CodeGraph indexed it in under 4 minutes.

MCP Tools

CodeGraph exposes 8 tools through its MCP server:

Category	Tool	Purpose
Search	codegraph_search	Find symbols by name using FTS5 full-text search
Search	codegraph_files	Get indexed file structure
Search	codegraph_status	Check index health and statistics
Navigation	codegraph_callers	Trace incoming call chains
Navigation	codegraph_callees	Trace outgoing call chains
Navigation	codegraph_node	Get symbol details with optional source code
Impact	codegraph_impact	Analyze what code is affected by changing a symbol
Context	codegraph_context	Build complete context for a task in one call

The codegraph_context tool is the most powerful — it combines search, navigation, and code retrieval into a single call, replacing the 40-50 tool calls that agents normally make during exploration.

Supported Languages

CodeGraph supports 19+ languages through tree-sitter grammar parsing:

Category	Languages
Web/Scripting	TypeScript, JavaScript, Python, Ruby, PHP, Dart, Svelte, Liquid
Systems	Go, Rust, C, C++
JVM	Java, Kotlin
Apple	Swift
.NET	C#
Other	Pascal/Delphi

Getting Started

Quick Install

npx @colbymchenry/codegraph

The interactive installer will:

1. Auto-detect installed agents (Claude Code, Cursor, Codex CLI, opencode, Gemini CLI, AntiGravity, Kiro, Hermes Agent)
2. Install codegraph on your PATH
3. Configure each agent’s MCP server settings and instructions
4. Set up auto-allow permissions for Claude Code
5. Initialize your current project

Non-Interactive Install

# Auto-detect agents, install global
codegraph install --yes

# Explicit target list
codegraph install --target=cursor,claude --yes

# Print config snippet without writing files
codegraph install --print-config codex

Initialize Projects

cd your-project
codegraph init -i

This builds the per-project knowledge graph index. Restart your agent for the MCP server to load.

CLI Reference

codegraph                         # Run interactive installer
codegraph install                 # Run installer (explicit)
codegraph init [path]             # Initialize in a project (--index to also index)
codegraph uninit [path]           # Remove CodeGraph from a project
codegraph index [path]            # Full index (--force to re-index)
codegraph sync [path]             # Incremental update
codegraph status [path]           # Show statistics
codegraph query <search>          # Search symbols (--kind, --limit, --json)
codegraph files [path]            # Show file structure
codegraph context <task>          # Build context for AI
codegraph affected [files...]     # Find test files affected by changes
codegraph serve --mcp             # Start MCP server

Affected Files for CI

# Pass files as arguments
codegraph affected src/utils.ts src/api.ts

# Pipe from git diff
git diff --name-only | codegraph affected --stdin

# Custom test file pattern
codegraph affected src/auth.ts --filter "e2e/*"

Library Usage

import CodeGraph from '@colbymchenry/codegraph';

const cg = await CodeGraph.init('/path/to/project');

await cg.indexAll({
  onProgress: (p) => console.log(`${p.phase}: ${p.current}/${p.total}`)
});

const results = cg.searchNodes('UserService');
const callers = await cg.getCallers('UserService.login');
const impact = await cg.getImpactRadius('UserService.login');
const context = await cg.buildContext('implement user authentication');

await cg.watch();  // Auto-sync on file changes
await cg.close();

Key Features Summary

Feature	Description
Token Optimization	Fewer tokens consumed — agents query structured graph instead of reading raw files
Smart Context Building	One tool call returns entry points, related symbols, and code snippets
Full-Text Search	Find code by name instantly across your entire codebase, powered by FTS5
Impact Analysis	Trace callers, callees, and the full impact radius of any symbol
Always Fresh	File watcher uses native OS events with debounced auto-sync
19+ Languages	TypeScript, JavaScript, Python, Go, Rust, Java, C#, PHP, Ruby, C, C++, Swift, Kotlin, Dart, Svelte, Liquid, Pascal/Delphi
Framework-aware Routes	Recognizes web-framework routing across 13 frameworks
100% Local	No data leaves your machine. No API keys. No external services
8+ Platform Support	Claude Code, Cursor, Codex CLI, opencode, Gemini CLI, AntiGravity, Kiro, Hermes Agent

What Changed Since Our Original Post

Aspect	Original Post (May 20)	This Update (June 4)
Stars	7,600	38,745
Tagline	“94% fewer tool calls, 77% faster exploration”	“fewer tokens, fewer tool calls, 100% local”
Platforms	4 (Claude Code, Cursor, Codex CLI, opencode)	8+ (added Gemini CLI, AntiGravity, Kiro, Hermes Agent)
Focus	Tool call reduction	Token optimization + tool call reduction
Diagrams	2 (architecture, features)	3 new (token flow, ecosystem, growth)

Conclusion

CodeGraph’s 5x growth in two weeks signals a shift in how developers think about AI coding agents. The bottleneck isn’t just tool calls — it’s tokens. Every unnecessary file read, every redundant grep search, every wasteful exploration loop consumes context window capacity that could be used for actual coding.

CodeGraph solves this by pre-indexing codebases into a semantic knowledge graph with tree-sitter, giving AI agents instant structural understanding that replaces dozens of file-scanning tool calls with a single graph query. With 19+ language support, 13 framework-aware route detectors, 8+ platform integrations, 100% local processing, and automatic file watching, CodeGraph is a zero-configuration productivity multiplier for any developer using AI coding agents.

Repository: https://github.com/colbymchenry/codegraph

npm: @colbymchenry/codegraph

License: MIT

Enjoyed this post? Never miss out on future posts by following us