RepoGraph - Repository Intelligence for Humans and AI

RepoGraph turns a codebase into a queryable intelligence graph for exploration, impact analysis, and architecture understanding.

It is a hybrid static + dynamic analysis tool:

• Static analysis maps files, symbols, imports, call edges, pathways, dead code signals, config usage, and invariants.
• Dynamic analysis overlays runtime traces (trace install + test/run + sync) to add observed evidence and correct static blind spots.

Core Capabilities

• Build a repository-wide symbol and call graph
• Generate pathway context docs for likely execution flows
• Classify dead code with confidence tiers and supporting signals
• Estimate blast radius with dependency and caller impact analysis
• Map module-level structure, config key usage, and architectural invariants
• Surface event topology, interface implementations, and constructor dependency hints
• Merge runtime traces into static analysis to improve confidence and reduce blind spots

Quick Start

./setup.sh
./run.sh

Or direct CLI:

pip install -e "."
repograph init
repograph sync --full
repograph summary

Requirements: Python 3.11+

Interfaces

CLI (full surface)

Primary interface for local development:

• index/sync: init, sync, status, watch, clean
• exploration: summary, report, modules, node, query, impact
• architecture: invariants, config, test-map, events, interfaces, deps
• pathways: pathway list, pathway show, pathway update
• runtime tracing: trace install, trace collect, trace report, trace clear
• integrations: mcp, export, doctor, test

Full command and flags: docs/CLI_REFERENCE.md

Python API

Use RepoGraph for scripts, test tooling, and service integrations:

from repograph.api import RepoGraph

with RepoGraph("/path/to/repo") as rg:
    rg.sync(full=True)
    print(rg.pathways())
    print(rg.dead_code())
    print(rg.full_report())

Surface details: docs/SURFACES.md

MCP Server (curated AI surface)

Expose RepoGraph to AI tools via MCP:

repograph mcp /path/to/repo

MCP intentionally exposes a curated subset of CLI/API methods.

Interactive Menu (guided full functionality)

RepoGraph includes a full interactive terminal menu:

repograph menu

The menu includes command browsing, plain-language explanations, and run presets so users can use the full tool without memorizing CLI flags.

Why use RepoGraph

• Reduce onboarding time in unfamiliar repositories
• Replace ad-hoc grep exploration with structured repository intelligence
• Improve change safety before refactors with impact-first analysis
• Give AI coding agents a reliable map of the project

Example workflow

repograph sync --full
repograph summary
repograph modules --issues
repograph pathway list
repograph pathway show <name>
repograph impact <symbol>

Then, if needed:

repograph trace install
pytest
repograph sync
repograph trace report

Documentation Map

• Start here: docs/README.md
• Setup and install tiers: docs/SETUP.md
• CLI flags and examples: docs/CLI_REFERENCE.md
• API/CLI/MCP boundaries: docs/SURFACES.md
• Pipeline and hooks: docs/PIPELINE.md
• Accuracy and known limits: docs/ACCURACY.md

License

RepoGraph is licensed under GNU AGPL v3.0.
See LICENSE.

---

Development Status

RepoGraph is under active development.
Interfaces and docs are being improved continuously; verify behavior against current CLI help and source when integrating deeply.