AI Code Review Agent

AI Code Review Agent

Multi-agent system powered by LangGraph. Analyzes pull requests for security, performance, style, and test coverage issues in parallel. Supports Groq (Llama 3.3), Gemini, OpenAI, and Anthropic.

LangGraph
Groq / Llama 3.3
Gemini
OpenAI
Anthropic
FastAPI
Review a Pull Request
Paste a GitHub PR URL to run a multi-agent code review. Typically takes 30-90 seconds.

No API key? Click Try Demo to see a real review of a freeCodeCamp PR.

🛡️

Security Agent

SQL injection, XSS, hardcoded secrets, auth flaws

⚡

Performance Agent

N+1 queries, memory leaks, algorithmic complexity

🎨

Style Agent

Naming, DRY, SOLID principles, readability

🧪

Test Coverage Agent

Missing tests, edge cases, assertion quality

Architecture

PR Diff
→
Smart Chunking
→
Security
Perf
Style
Tests
→
Aggregator
→
Meta Reviewer
→
Report

4 specialist agents run in parallel via LangGraph fan-out, then converge through aggregation and meta-review