https://github.com/basith-ahmed/git-ask
an AI-powered GitHub repository analysis and RAG platform using Langchain, Supabase, PostgreSQL, TRPC, Next.js, Gemini, and GitHub API.
https://github.com/basith-ahmed/git-ask
gemini langchain nextjs postegresql t3-stack trpc
Last synced: about 2 months ago
JSON representation
an AI-powered GitHub repository analysis and RAG platform using Langchain, Supabase, PostgreSQL, TRPC, Next.js, Gemini, and GitHub API.
- Host: GitHub
- URL: https://github.com/basith-ahmed/git-ask
- Owner: basith-ahmed
- Created: 2025-02-04T17:35:22.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-06-27T05:49:16.000Z (12 months ago)
- Last Synced: 2025-06-27T06:30:40.686Z (12 months ago)
- Topics: gemini, langchain, nextjs, postegresql, t3-stack, trpc
- Language: TypeScript
- Homepage: https://ask-git.vercel.app
- Size: 343 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Git-Ask
## Overview
Git-Ask is an application that allows users to submit a GitHub repository URL and ask questions related to that repository. The application fetches repository contents, generates embeddings for each file, and stores them in a Supabase database. Users can then query the repository and receive responses based on the stored embeddings.
## Technologies Used
* **Frontend**: Next.js
* **Backend**: Langchain, trpc for type-safe APIs
* **Database**: Supabase (PostgreSQL)
* **GitHub API**: Used for fetching repository contents
## Features
* **GitHub Repo Integration**: Users can submit a GitHub repository URL.
* **Embedding Generation**: Backend fetches repository contents, generates embeddings, and stores them.
* **Query Processing**: Backend processes user queries and generates relevant responses based on stored embeddings.
## How It Works
1. **Repository Submission**: Users enter a GitHub repository URL.
2. **Backend Processing**:
* The backend fetches contents from the GitHub repository.
* Embeddings are generated for each file using Langchain and stored in Supabase.
3. **Query and Response**:
* Users can ask questions related to the repository.
* Queries are sent to the backend, processed using trpc APIs, and responses are generated based on stored embeddings.