Your codebase as a graph.
Every issue visible.

Repotoire builds a knowledge graph of your code—combining AST analysis, semantic understanding, and graph algorithms—to find architectural debt, code smells, and issues linters miss.

8 hybrid detectors·10-100x faster re-analysis·AI-powered auto-fix

myapp/Last scan: 2 min ago

Health Score72

Structure (40%)

Quality (30%)

Architecture (30%)

3 circular dependencies

12 architectural bottlenecks

847 dead exports

23 modularity issues

156 code quality findings

Linters see files. We see relationships.

Traditional tools analyze files in isolation. Repotoire builds a Neo4j knowledge graph combining structural, semantic, and relational analysis.

	Traditional Linters	Repotoire
Analysis scope	Single file at a time	Entire codebase as graph
Architectural issues	Not detected	Cycles, bottlenecks, modularity
Dead code detection	Unused variables only	Cross-file export analysis
Re-scan time (10k files)	Full rescan: 5-15 min	Incremental: 8 seconds
Fix generation	Text suggestions	GPT-4o + RAG code diffs
Health scoring	Pass/fail per rule	Structure + Quality + Architecture

8 hybrid detectors

Graph algorithms + traditional tools working together. Each detector integrates Ruff, Pylint, Mypy, Bandit, or Semgrep with cross-file graph analysis.

Circular Dependencies

Detects import cycles across any depth using Tarjan's algorithm on the dependency graph.

A → B → C → A

Dead Code

Finds exports never imported anywhere. Cross-file analysis, not just single-file unused vars.

847 unused exports

Bottleneck Analysis

Identifies high-fanin modules that cause cascading test failures when changed.

utils.ts → 234 dependents

Modularity Index

Measures cohesion within and coupling between modules. Flags tightly coupled packages.

Q = 0.42 (target: 0.7)

Code Smells

Integrates Ruff + Pylint findings with semantic context from the knowledge graph.

Long method, god class

Type Coverage

Mypy integration with cross-module type flow analysis. Finds untyped boundaries.

78% typed (12 gaps)

Security Scan

Bandit + Semgrep rules enhanced with data flow analysis through the graph.

SQL injection path found

Complexity Hotspots

Combines cyclomatic complexity with change frequency from git history.

payment.py: high churn + complexity

AI-powered auto-fix

Every issue comes with a GPT-4o generated fix using RAG over your codebase. 70%+ approval rate in production. Apply with one click or export as PR.

Setup in 5 minutes

Connect your repo, build the graph, get actionable insights.

Connect

# Install CLI

npx repotoire init

✓ Connected to github.com/acme/app

✓ Found 2,847 files

Analyze

Building knowledge graph...

Nodes: 12,458

Edges: 47,921

Fix

Circular dependency

auth.ts → user.ts → auth.ts

How we compare

Traditional

Repotoire

Speed

5-15 min full codebase scans

8 second incremental analysis

Detection

Single-file syntax issues

Cross-file architectural problems

Fixes

Generic suggestions

AI-generated code with 70% approval rate

Regex pattern matching

Natural language queries

Results from real teams

circular dependencies fixed

in a 200k LOC TypeScript monorepo

12 min → 8 sec

CI time reduction

with incremental analysis

2,340

dead exports removed

saving 18% bundle size

"We found circular dependencies that had been silently breaking hot reload for 2 years. Fixed them all in an afternoon."

— Engineering Lead at a Series B startup (200k LOC codebase)

Pricing

Free forever for personal projects.

Free

For individual developers

$0/dev/mo

1 repository
All 8 detectors
HTML reports
CLI access

Team

For growing teams

$79/dev/mo

Unlimited repos
AI auto-fix
GitHub Actions
Natural language search
Slack alerts

Enterprise

For large organizations

Custom

Self-hosted option
SSO/SAML
Custom detectors
Dedicated support

FAQ

Join 500+ engineering teams

Stop shipping bugs linters miss

Find architectural issues in seconds. Fix them with AI. Start free.

No credit card required. 14-day free trial.

Your codebase as a graph.Every issue visible.