Repository context
See what your AI setup is missing.
Start with a public repository preview. Inspect the findings, then sign in to save an analysis and prepare context files.
What you get — worked example
acme/storefrontVUE 3 + SPRING BOOTanalyzed in 41s
AI-context health
87
Frameworks detected
8
Conventions profiled
14
Security findings
1
Detected stack & conventions
Vue 3 (Composition API)PiniaViteTypeScriptSpring Boot 3Spring Data JPAPostgreSQLMaven
Existing CLAUDE.md scored 61/100 — stack section accurate; conventions drifted since March. A merge-safe upgrade is ready.
Unpinned MCP server — .mcp.json references a server without a version pin — an agent-supply-chain risk.
Suggested agent roster — frontend-reviewer, api-contract-guard, db-migration-checker — grounded in your module layout.
Illustrative — your report references your actual frameworks, types, and conventions, not these placeholders. The real analysis goes much deeper.
Token estimator
How big is your codebase?
Point it at a folder (or pick files) for a rough token estimate. Bigger codebases use more tokens per analysis. Nothing is uploaded — this runs entirely in your browser.
For reference
A small library / script~5k tokens
A typical web app~50–150k tokens
A large monorepo500k+ tokens
Rough estimate (~4 characters per token, derived from file sizes). Actual usage depends on which files are analyzed and how deep the analysis goes.
Then keep it healthy — automatically
Pro adds the GitHub App: health checks on every PR, daily drift monitors, and one-click setup PRs across up to 25 repos.