ResearchAI is a state-of-the-art Research Assistant platform designed to streamline academic workflows. Leveraging High-Performance RAG (Retrieval-Augmented Generation), Groq-powered Llama 3, and Fast Vector Search, it empowers researchers to synthesize, analyze, and interact with academic papers at scale.
- Centralized Management: Seamlessly upload, tag, and organize your research library.
- Paper Indexing: Automatic text extraction and vectorization for instant retrieval.
- Semantic Search: Find specific concepts across your entire collection, not just keywords.
- Multi-Paper Context: Ask questions across a single paper or your entire selection.
- Confidence Scores: Transparency in AI reasoning with real-time groundedness metrics.
- Source Attribution: Pinpoint exactly where in the paper the AI found the answer.
- Streaming Responses: Real-time insights via Socket.io for a near-instant experience.
- Literature Reviews: Generate comprehensive syntheses across multiple studies in seconds.
- Structural Summaries: Instant breakdowns of Problem, Approach, Methodology, Results, and Contribution.
- Comparison Engine: Analyze methodological differences and result variations side-by-side.
- Sticky Note Annotations: Highlight key passages and attach your own research notes.
- AI Flashcards: Automatically generate study aids based on complex paper concepts.
- Export Center: One-click exports to PDF, BibTeX, APA, and MLA formats.
| Layer | Technologies |
|---|---|
| Frontend | HTML5, Tailwind CSS, Vanilla JavaScript, Socket.io, PDF.js |
| Backend | Node.js (Express), SQLite3 (Auth), Neo4j (Graph References) |
| AI / LLM | Groq SDK (Llama 3 70B), FAISS (Vector Store), Python (FastAPI) |
| DevOps | Railway, Docker / Nixpacks |
├── project-root/ # Main Node.js Express server
│ ├── app.js # Core application logic
│ ├── routes/ # API endpoints (Auth, Library, AI, Export)
│ ├── frontend/ # Web assets (HTML, CSS, JS)
│ └── scripts/ # Background processing & embedding tools
├── faiss-service/ # Python microservice for vector operations
│ └── main.py # FastAPI frontend for FAISS
└── README.md # You are here
Create a .env file in project-root/ based on .env.example:
GROQ_API_KEY="your_groq_key"
NEO4J_URI="bolt://..."
NEO4J_USER="neo4j"
NEO4J_PASSWORD="..."
JWT_SECRET="..."The Python service handles high-speed similarity search for RAG.
cd faiss-service
pip install -r requirements.txt
uvicorn main:app --port 8000cd project-root
npm install
npm run build # Compiles Tailwind CSS
npm startAccess the platform at http://localhost:3000
This project is licensed under the ISC License.
Powered by Groq and Llama 3.