Git Intel is an advanced intelligence platform that provides deep insights into GitHub organizations and repositories. It leverages AI-driven analytics to assess authority levels, track project involvement, and summarize key discussions using state-of-the-art machine learning models.
A detailed overview of the architecture involves the following
Git Intel is built using the T3 Stack, which includes:
- Next.js 15 - Server-side rendering and frontend framework
- TypeScript - Strongly-typed language for reliability
- tRPC - End-to-end type-safe API communication
- Prisma - ORM for database interactions
- TailwindCSS - Utility-first CSS framework for rapid UI development
- ShadCN - Modern UI components built on Radix and Tailwind
- NeonDB - Serverless Postgres database (also used for vector storage)
- GitHub OAuth - Authentication provider for user login and permissions
- Gemini AI (via API) - Used for Retrieval-Augmented Generation (RAG) insights
- Gemini Embeddings (768 dimensions) - To power search and AI-driven analytics
- AssemblyAI - Automated meeting note generation and summarization
- Firebase - Authentication and real-time database functionalities
-
GitHub Organization Insights - Determine the authority level of users and analyze team structures.
-
Project Tracking - Identify key contributors and active repositories.
-
RAG-Powered AI Queries - Leverage Gemini AI for intelligent query responses based on GitHub data.
-
Vector Search - Store and retrieve embeddings efficiently using Neon.tech's vector capabilities.
-
Automated Meeting Summaries - Use AssemblyAI to transcribe and summarize discussions.
-
Modern UI - Built with ShadCN for a sleek and responsive experience.
Contributions are welcome! Feel free to submit a pull request or open an issue.
This project is licensed under the MIT License.
💡 Git Intel – Empowering developers with AI-driven GitHub intelligence.