Understand Any Codebase in Minutes.
AI-powered repository intelligence for engineering teams. Analyze architecture, uncover technical debt, review pull requests, and onboard developers faster.
active-repository-scan
github.com/codeforge-ai/production-serviceAutonomous Repository Intelligence
CodeForge AI analyzes repositories to uncover architecture, security risks, technical debt, and code insights using AI-powered repository intelligence.
Full Stack Engineering Intelligence
Everything you need to scan, audit, and understand your codebase.
How CodeForge AI Works
Connect
Authenticate securely using GitHub OAuth.
Import
Select a public repository or paste a clone URL.
Analyze
FastAPI clones and parses symbols in background.
Index
Gemini generates 768-dimensional RAG embeddings.
Explore
Interact with graphs, chat RAG, and audit metrics.
Index supported source files
Built With Modern Enterprise Technology
CodeForge AI leverages best-in-class tools and infrastructure to deliver instant codebase intelligence.
Explore the Core Console
Architecture Visualizer
View local imports mapped to a React Flow canvas with animated connection lines.
Codebase Chat (RAG)
Ask detailed logical questions and see responsive answers with file citation tokens.
Security Audits
Inspect vulnerability summaries, severity badges, and file locations instantly.
Health Scores & Grades
Inspect measured source metrics, confidence, and explicitly unmeasured dimensions.
Deployment options — paid plans are not available
Frequently Asked Questions
Repositories are cloned into temporary directories and cleaned up after analysis. Source chunks, embeddings, reports, and chat messages are persisted in Supabase. Retrieved chunks are sent to the configured AI providers. Configure ownership policies and retention before handling private code.
Supported source files are indexed as text for repository chat. Import analysis varies by language. Deterministic security checks currently cover selected secret patterns and Python source rules; other languages do not receive equivalent security coverage.
Private repositories require a GitHub OAuth token with access to that repository. Configure GitHub scopes, Supabase ownership policies, and server access controls. Scans do not execute repository code; production deployments still need process isolation.
You can run the frontend and backend using the setup instructions in the repository, with your own Supabase, Gemini, and Groq configuration. Review the repository license and deployment limitations before use.
Get in Touch
Have a question or looking for a custom enterprise setup? Send us a message.