Upload any documents. Ask anything. Get accurate answers with citations. Built for the Google Cloud Rapid Agent Hackathon — Elastic Track.
- Upload multiple documents (PDF, TXT, DOCX, CSV, MD)
- Ask questions in plain English — no commands, no syntax
- Get precise answers with exact source citations (document + page/offset number)
- Reason across multiple documents simultaneously
- Multi-turn conversation — remembers context of your questions
User → Streamlit UI → FastAPI Backend → Hybrid Search (Elastic) → Gemini 2.5 Flash Lite → Answer with Citations
| Component | Technology |
|---|---|
| AI Model | Google Gemini 2.5 Flash Lite |
| Embeddings | Gemini text-embedding-004 |
| Search Engine | Elasticsearch (Elastic Cloud) |
| MCP Integration | Elastic Agent Builder MCP Server |
| Backend | FastAPI + Python |
| Frontend | Streamlit |
| Backend Host | Render.com |
| Frontend Host | Streamlit Cloud |
- Law firms → Search across hundreds of case files instantly
- Hospitals → Query patient records, policies, and guidelines
- Universities → Students search through research papers and notes
- Companies → Employees find answers in internal documents
# Clone the repository
git clone https://github.com/adhipatya3552/SearchBot-Agent.git
cd SearchBot-Agent
# Set up local environment variables
cp .env.example .env
# Open .env and fill in your real keys
# Run the FastAPI backend
cd backend
pip install -r requirements.txt
uvicorn main:app --reload
# In a separate terminal, run the Streamlit frontend
cd ../frontend
pip install -r requirements.txt
streamlit run app.py- SaaS subscription: ₹499/month per organization
- API access for enterprises
- White-label licensing
Apache 2.0 — see LICENSE