Open Code Review is a free, open-source CLI that catches AI-specific code defects invisible to traditional linters like ESLint and SonarQube. **The problem**: AI coding assistants (Copilot, Cursor, Claude) generate code that looks correct but contains subtle defects — hallucinated package imports, stale APIs from training cutoffs, logic contradictions across files, and unnecessary abstractions. Traditional linting tools miss all of these. **How it works**: Deep Scan combines three analysis layers: 1. **Structural Detection** — AST analysis catches hallucinated imports by cross-referencing npm/PyPI/Maven registries, detects stale APIs, security anti-patterns, and over-engineering 2. **Semantic Analysis** — Embedding similarity recall identifies context window artifacts and ranks risk scoring 3. **LLM Deep Scan** — Sends suspicious code to a local LLM (Ollama) or any OpenAI-compatible provider for cross-file coherence checks and confidence scoring **Why it's different**: - 100% free and open-source (BSL 1.1 → Apache 2.0 in 2030) - Works with 6 languages (TypeScript, Python, Java, Go, Kotlin, JavaScript) - 100% local option with Ollama — no code leaves your machine - GitHub Action + GitLab CI integration in 30 seconds - AI auto-fix with `ocr heal` command - SARIF output for GitHub Code Scanning **Quick start**: `npm i -g @opencodereview/cli && ocr scan src/` GitHub: https://github.com/raye-deng/ope... Portal: https://codes.evallab.ai

Open Code Review

Catch AI-generated code defects linters miss

Catch AI-generated code defects linters miss