polycodegraph

pip install polycodegraph # the PyPI distribution name codegraph init # the CLI binary + Python module + MCP server are all codegraph (see footnote ↓) codegraph build # parse repo → .codegraph/graph.db codegraph serve # web dashboard at http://127.0.0.1:8765

That's it. Three commands and you have a queryable graph, a 3D dashboard, and an MCP server your IDE can talk to.

Adding a new language is a single tree-sitter parser module + fixture file (~3 hours — seecodegraph/parsers/go.pyfor the v1 template). PRs welcome.

polycodegraph has exactly one opinion:build the right graph, and every interesting feature falls out for free.

The inputs that feed the graph go beyond imports and call edges. polycodegraph reads tree-sitter parses forPython, TypeScript, JavaScript, and Go; capturesevery call-site's arguments as text; recognizes24 framework decoratorsso FastAPI / Flask / Celery / pytest / Click / MCP / Django / SQLAlchemy handlers are never confused with dead code; detectsroutes(@app.get("/x")) andfrontend fetches(fetch,axios,useSWR,useQuery); andstitches URLs across the stack(/{id} ↔ ${id} ↔ :id) so it can trace a fetch all the way to its handler.

The outputs that comefor freeonce the graph is right:

Decorator-aware dead code, role classification (HANDLER / SERVICE / COMPONENT / REPO), blast radius, cycles, untested-function detection, an end-to-end cross-stack trace with rename annotations, a 3D focus-mode dashboard, a Learn Mode lifecycle modal, local embeddings for semantic + hybrid search, an 18-tool MCP server, and a PR-review CI that graph-diffs the branch againstmain.

One SQLite file. No daemon. No network. Travels with your git branch.

┌─────────────────────────────────────────────────────┐ │ tree-sitter parsing │ │ (Python, TS/JS, TSX, JSX, Go) │ └─────────────────────────────────────────────────────┘ ↓ ┌─────────────────────────────────────────────────────┐ │ Cross-file resolution (R1, R2, R3) │ │ ✓ per-name imports ✓ relative imports │ │ ✓ constructor calls ✓ decorators │ │ ✓ self.X.Y chains ✓ fresh instances │ └─────────────────────────────────────────────────────┘ ↓ ┌─────────────────────────────────────────────────────┐ │ SQLite graph (nodes + edges) │ │ DF0: call-site arguments │ │ DF1: routes (FastAPI, Flask, aiohttp) │ │ DF2: fetches (fetch, axios, SWR, useQuery) │ │ DF3: URL stitching (/{id} ↔ ${id} ↔ :id) │ │ DF4: end-to-end trace (fetch→handler→service→DB) │ └─────────────────────────────────────────────────────┘ ↙ ↓ ↘ CLI tools Web dashboard MCP server (graph, roles, (3D focus view, (18 tools for cycles, dead architecture, Claude Code, code, untested) learn mode) Cursor, etc.)

Benchmark — same Claude, varying graph MCP

Four configurations. Same Claude Sonnet 4.6. Same 10 questions across two real codebases (polycodegraph itself + FastAPI).All four configs include Claude's native grep + file-reading tools— what every dev gets out of the box in Claude Code or Cursor. The only thing that changes is whether a graph MCP isalsoregistered alongside.

- claude+grepalone is the most correct (8/10)— Claude can answer most codebase questions by grepping and reading whole files. But it pays the price:336k tokens, $1.17, 78s avg latency.
- + polycodegraphmatches that within one question (7/10) at3× lower cost and 4× lower latency(90k tokens, $0.37, 20s).Because polycodegraph returns small focused subgraphs (~20-50 tokens per call) instead of grep-dumping whole files into Claude's context.
- The other graph MCPs are strictly worse than just grepping.code-review-graph: 3/10 at $0.68. graphify: 5/10 at $0.50. They add tool overhead without paying off in correctness.

pip install polycodegraph codegraph init codegraph build

codegraph initwrites a project-level.mcp.jsonin the repo —Claude Code and Cursor auto-pick that upas soon as you open the project. For other clients you currently need to add the server to their global config manually (v0.2 will do this for you).

- Type inference(Mypy / Pyright). DF0 captures argumenttext, not types. Roadmap v0.3.
- Argument-value identity across hops.DF4 emits ordered hops with rename annotations; full single-value propagation from fetch body → route param → service arg → DB column is deferred (v0.3).
- Docstrings are stored on every node but not yet consumed by analysis.Embeddings use them as fallback body text; dead-code, role classification, and dataflow ignore them. Roadmap v0.3.
- Git-history mining(commit-message semantics, author / touch-frequency signals). Not implemented. Git is used only for the current HEAD SHA and PR-review diff. Roadmap v0.4.
- Per-language resolver parity(v0.1.2). Python ships the full R1/R2/R3 fixes. TypeScript R2 patterns (path aliases, fresh-instance binding, decorator-call edges) are deferred.
- Typer CLI symbols are not tagged HANDLER(v0.1.x). DF1.5 only classifies HTTP framework decorators.
- Async / await visualization(v0.4). DF4 walks the synchronous call graph only.
- Error-path branch rendering(v0.4). Learn Mode shows the happy path.
- Auth middleware as a distinct phase(v0.4). Today auth shows up as a regular CALL node.
- Multi-param simultaneous highlighting(v0.4). Single-param selection only.
- Cross-process traces(v0.4). Can't yet link multiple.codegraph/graph.dbfiles.

We run polycodegraph on its own source as a regression target. Dead-code findings dropped from451 → 24+ → 15 → 0as the resolver hardened, decorator-aware entry-point detection landed, and intentional public-API methods were marked with# pragma: codegraph-public-api.

- 3,320 nodes(files, classes, functions, imports)
- 7,557 edges(5,245 CALLS, 1,357 DEFINED_IN, 886 IMPORTS, 28 INHERITS, 12 ROUTE, 27 FETCH_CALL, 1 READS_FROM, 1 WRITES_TO)
- 3 cycles, all documented and accepted (dashboard redraw, parser self-recursion, MCP serve/run resolver false positive)
- 0 dead-code findings(with pragma exemptions for public-API methods)
- 637 tests passing(537 Python pytest + 100 Node tests)

The wedge isn't a fancier graph algorithm — it's that polycodegraph treatstrace this argument across the stackas a first-class operation, not a follow-up grep. Embedding-based retrieval tools (code-review-graph, Cursor, Cody) handle prose / docstrings well; the right architecture isgraph + embeddings in the same MCP loop, and v0.1.0 ships both.

This project is installed from PyPI aspolycodegraphbecause the bare namecodegraphwas already taken when v0.1.0 shipped. Everything else — the Python package you import, the CLI binary you run, and the MCP server key you register — iscodegraph, the original project name. We're planning to unify onpolycodegrapheverywhere in v0.2 (CLI rename with acodegraphalias for one release). For now: two names, one tool.

This is a web browser that enables your coding agent, such as Claude Code, to visit websites on your behalf and assist you in identifying bugs or creating UI test cases.

A stateful LSP runtime for AI agents: warm language server sessions with 50+ tools for go-to-definition, find-references, diagnostics, rename, and more across 30+ languages.

MCP server that gives coding agents program-analysis primitives — data flow, call graphs, taint analysis — so they reason from ground truth instead of grep-and-guess. (same as the GitHub About — keeps your messaging consistent across the web).

An MCP service that equips your workspace with a complete set of AI-accessible development tools for reading, editing, executing, and managing code.

Visual Studio extension with 20 Roslyn-powered MCP tools for AI assistants. Semantic code navigation, symbol search, inheritance, call graphs, safe rename, build/test.

A tool for safely executing local Python code without requiring external data files.

AI-to-AI code review platform — Claude, Codex, and Gemini cross-check each other via MCP, REST API, and CLI for consensus-based results.

Persistent code index using Tree-sitter for fast, precise code search. Replaces grep with ~50 token responses instead of 2000+.

AI-powered code quality analysis to detect best practice violations, security issues, and architectural problems in real-time.

Orchestrates a dual-AI engineering loop where a Primary AI plans and implements, while a Review AI validates and reviews, with continuous feedback for optimal code quality. Supports custom AI pairing (Claude, Codex, Gemini, etc.)

AmazingMCP — MCP Server for .NET / C# Codebases