Skip to main content

​
Starting the application

RepoRAGX runs as an interactive CLI tool. To start it, run: 
python -m src.main

​
Welcome banner

When you start RepoRAGX, you’ll see the ASCII art banner: 
/**
 *    __________                    __________    _____    ____________  ___
 *    \______   \ ____ ______   ____\______   \  /  _  \  /  _____/\   \/  /
 *     |       _// __ \\____ \ /  _ \|       _/ /  /_\  \/   \  ___ \     / 
 *     |    |   \  ___/|  |_> >  <_> )    |   \/    |    \    \_\  \/     \ 
 *     |____|_  /\___  >   __/ \____/|____|_  /\____|__  /\______  /___/\  \
 *            \/     \/|__|                 \/         \/        \/      \_/
 */

Chat with your github repository

​
Interactive setup flow

After the banner, RepoRAGX will prompt you for the required information to connect to your repository and start the RAG pipeline.
1

Enter GitHub token

The first prompt asks for your GitHub Personal Access Token. This is hidden for security:
GitHub Personal Access Token: ********
The token needs content:read permission to access repository files. Get yours at github.com/settings/tokens.
2

Enter Groq API key

Next, you’ll be prompted for your Groq API key (also hidden):
Groq API Key: ********
Get your API key from console.groq.com/keys.
3

Select LLM model

Choose which Groq model to use for answering questions:
Model Name (default: llama-3.3-70b-versatile, see Groq docs for supported models):
Press Enter to use the default llama-3.3-70b-versatile model, or specify a different model supported by Groq.
4

Specify repository

Enter the repository you want to chat with:
Repo (owner/repo): AnmolTutejaGitHub/RepoRAGX
Use the format owner/repo (e.g., facebook/react, vercel/next.js).
5

Select branch

Choose which branch to index:
Branch (default: main): main
Press Enter to use main, or specify another branch like develop or staging.

​
Loading and indexing process

Once you’ve provided all inputs, RepoRAGX starts the data ingestion pipeline: 
Initilizing github loader.....
Fetching files from github....
Loaded 42 files from github!
Splitting documents into chunks...
chunking completed
Initilizing RAG Retriver pipeline
Initializing Groq LLM...
The indexing process downloads all repository files (excluding binaries and common folders like node_modules/), chunks them using language-aware splitting, generates embeddings, and stores them in ChromaDB.

​
What gets loaded

RepoRAGX filters out files that aren’t useful for code understanding:
  • Excluded file types: Images (.png, .jpg, .svg), archives (.zip, .tar), binaries (.exe, .dll), media files (.mp3, .mp4), and more
  • Excluded folders: node_modules/, .git/, dist/, build/, __pycache__/, venv/, .venv/

​
Storage location

Vector embeddings are persisted locally in: 
~/.RepoRAGX/vector_store/<owner>_<repo>/
This means you only need to index a repository once. Future runs will reuse the existing embeddings.

​
Ready to chat

Once indexing completes, you’ll see the main query prompt: 
Ask anything ('exit' to quit):
You’re now ready to ask questions about the codebase. See the Querying repositories guide for details on how to interact with your indexed repository.

​
Exiting the application

To exit RepoRAGX at any time, type exit at the query prompt: 
Ask anything ('exit' to quit): exit
The application will terminate and return you to your shell.

Build docs developers (and LLMs) love

Get started for freeTalk to us