Revio

Next.jsReactTypeScriptPrisma ORMPostgreSQLOpenAI APIGemini APIQdrantBullMQRedisTree-sitterBabel ParserMermaidZodTailwindCSSZustandOctokit

Production-ready SaaS that reviews pull requests using Tree-sitter-powered code graphs and vector retrieval (RAG). It generates 1-5 confidence scores for merge readiness, detects security risks, visualizes blast radius, and adapts to team feedback over time.

Revio screenshot

About the Project

Revio is an AI code review platform that combines Tree-sitter parsing, code graph analysis, and RAG retrieval to deliver context-aware PR feedback. It helps teams catch bugs and security issues before merge, understand impact, and improve review quality over time.

Key Features

  • Tree-sitter + AST-powered code graph analysis with function relationships and impact radius
  • 1-5 confidence scoring for merge readiness
  • Adaptive learning system that reduces low-value review noise from team feedback
  • Interactive @revio-bot conversations in PR comments
  • Blast radius visualization with risk assessment
  • Vector-based semantic code search (embeddings + Qdrant)
  • Automated PR reviews via GitHub webhooks with inline comments
  • Security scanning (OWASP/CWE, secrets detection, SQLi, XSS)
  • AI-generated docstring suggestions and sequence diagrams
  • Auto-detection of coding standards (.claude.md, .cursorrules, agents.md)
  • Multi-tenant organizations with role-based access
  • Incremental indexing with encrypted token storage
Revio | Mayur Bijarniya