Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MIRA-CORA — Self-Challenging Multilingual RAG Assistant

An AI assistant that doesn't just answer — it questions its own answers before giving them to you, with retrieval-augmented, evidence-backed responses in the user's own language.

What It Does

Most RAG assistants retrieve, generate, and stop. MIRA-CORA adds a self-challenging step: it critiques its own draft answer against the retrieved evidence before responding, which reduces confident-but-wrong answers and surfaces a confidence score alongside each response.

Features

  • Self-Challenging AI — actively questions and re-checks its own answers against retrieved evidence
  • Multilingual Support — responds in whichever language the user writes in
  • RAG Integration — evidence-aware responses with confidence scoring, not just generation
  • Real-time Chat — fast, responsive interface
  • Graceful Error Handling — clear fallbacks when retrieval or generation fails

Tech Stack

Frontend: Vite, React, TypeScript, shadcn-ui, Tailwind CSS Backend: Python, OpenAI API

API

  • POST /ask — main chat endpoint
  • GET / — health check

Getting Started

Prerequisites: Node.js v18+, Python 3.11+, an OpenAI API key

# Backend
cd backend
pip install -r requirements.txt
python main.py

# Frontend (in a new terminal)
npm install
npm run dev

Add your OpenAI API key in backend/main.py before running.

App runs at http://localhost:8080.

Build for Production

npm run build

Build artifacts are output to dist/.

License

MIT — free to use for personal or commercial purposes.

About

Self-challenging multilingual RAG assistant with evidence-based confidence scoring

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages