Skip to main content

Git-aware code knowledge graph — incremental AST indexing on every commit, MCP server for AI assistants

Project description

gitcortex

A code knowledge graph for Git repositories. GitCortex indexes your codebase on every commit and answers questions like "what calls this function?" or "what changed between these two branches?" — so your AI coding assistant can work with real structure instead of scanning raw files.

Install

pip install gitcortex
# or
pipx install gitcortex
# or
uv tool install gitcortex

Supports macOS (Apple Silicon + Intel), Linux (x86_64 + arm64), and Windows (x64). No Rust or compiler required — a pre-built binary is bundled in the wheel.

Quick start

cd your-repo
gcx init       # index the repo + install git hooks
gcx serve      # start the MCP server for your AI assistant

After gcx init, the graph updates automatically on every git commit, merge, rebase, and checkout. No manual re-runs needed.

What it does

GitCortex builds a queryable graph of your codebase — functions, structs, classes, interfaces, call relationships, inheritance chains — and keeps it current automatically.

  • Works with: Rust, Python, TypeScript, JavaScript, Go, Java
  • Integrates with: Claude Code, Cursor, Windsurf, GitHub Copilot, Antigravity
  • Per-branch graphs: switching branches gives you the graph for that branch instantly
  • Zero runtime dependency: single self-contained binary, nothing else to install

Commands

gcx init

Index the current repo, install git hooks, and register the MCP server with your editor.

gcx init                     # auto-detects your editor
gcx init --editor cursor     # target a specific editor: claude, cursor, windsurf, copilot
gcx init --editor all        # write configs for every supported editor

gcx serve

Start the MCP server so your AI assistant can query the knowledge graph.

gcx serve

Once running, your AI assistant has access to tools like find_callers, lookup_symbol, list_definitions, trace_path, and more.

gcx query

Query the graph from the terminal without an AI assistant.

gcx query lookup-symbol MyStruct
gcx query find-callers process_request
gcx query list-definitions src/lib.rs

gcx blast-radius

See which parts of the codebase are affected by changes between two branches — useful before merging.

gcx blast-radius --base main --head feat/my-feature
gcx blast-radius --base main --head feat/my-feature --format github-comment

gcx viz

Open an interactive graph in your browser.

gcx viz                         # opens on port 5678
gcx viz --port 9000
gcx viz --branch feat/my-feature
gcx viz --format dot > graph.dot   # export as Graphviz DOT

gcx export

Generate a readable Markdown map of the codebase at .gitcortex/context.md. The git hook keeps it up to date automatically.

gcx export

gcx status

Show node and edge counts for the current branch.

gcx status

gcx clean

Wipe the local graph store and re-index from scratch on the next commit.

gcx clean

MCP tools available to your AI assistant

Tool What it answers
lookup_symbol Where is MyStruct defined?
find_callers What calls process_request?
find_callees What does handle_request call?
list_definitions What's defined in src/auth.rs?
find_implementors What implements AuthProvider?
trace_path How do you get from main to validate_token?
find_unused_symbols What's never called (dead code candidates)?
get_subgraph Everything within 2 hops of UserService
detect_changes What changed + who's affected vs main?
symbol_context Callers, callees, and uses for a symbol

All tools accept an optional branch parameter.

Configuration

.gitcortex/config.toml (committed to your repo, shared with your team):

[index]
languages = ["rust", "typescript", "python", "go"]
max_file_size_kb = 500

.gitcortex/ignore (.gitignore syntax — files to exclude from indexing):

target/
build/
**/*.generated.rs

License

MIT — free for commercial and open-source use.

GitHub · Issues

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

gitcortex-0.3.0-py3-none-manylinux2014_x86_64.whl (11.1 MB view details)

Uploaded Python 3

gitcortex-0.3.0-py3-none-manylinux2014_aarch64.whl (9.7 MB view details)

Uploaded Python 3

gitcortex-0.3.0-py3-none-macosx_11_0_arm64.whl (7.9 MB view details)

Uploaded Python 3macOS 11.0+ ARM64

gitcortex-0.3.0-py3-none-macosx_10_12_x86_64.whl (7.9 MB view details)

Uploaded Python 3macOS 10.12+ x86-64

File details

Details for the file gitcortex-0.3.0-py3-none-manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for gitcortex-0.3.0-py3-none-manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 c40fe11045d48c388cce34d9b366aacd36abaa1662974206d38ebb94900c7012
MD5 77c1392d60384bc1b7fb84be399e85d0
BLAKE2b-256 f2e8981e8b18ceb92bc57c0e8ec161ed75779b4749f222e4a9a7e94d52a88f95

See more details on using hashes here.

File details

Details for the file gitcortex-0.3.0-py3-none-manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for gitcortex-0.3.0-py3-none-manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 7e1d23d5c287edf26f1bb42938f21e48dffd8d7c52be3e1b1c5c186276912bcd
MD5 bdf704886db617afb0a3f9948e434f5b
BLAKE2b-256 fbf5c20d7d13620100ff014552bb13c8b5f2f07b6e9811ab965f98fe54135490

See more details on using hashes here.

File details

Details for the file gitcortex-0.3.0-py3-none-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for gitcortex-0.3.0-py3-none-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 6592da142d0231214d174a92214b8fc50d9efb51d1dfeccd2d0cf325e56963fd
MD5 244c2c1520e4db8e8bf0da9f5c1766df
BLAKE2b-256 630e6f289b1ec9e13ad0f4a22c5b4fe1b6cae34757b26058017eeb7466027a5a

See more details on using hashes here.

File details

Details for the file gitcortex-0.3.0-py3-none-macosx_10_12_x86_64.whl.

File metadata

File hashes

Hashes for gitcortex-0.3.0-py3-none-macosx_10_12_x86_64.whl
Algorithm Hash digest
SHA256 523d44ed32dcc7bff34946776114dbcb37ddafd0521a2150de64dbc08ef82bf2
MD5 d4605bdf943464d74367618119da6f14
BLAKE2b-256 14abea7b244b8ccec1746925feb42d3a190dd056a289d1c06526136a508d679b

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page