Talk to Data

A production-ready Text-to-SQL agent that converts natural language questions into SQL queries. Connect it to your database, ask questions in plain English, and get instant answers with visualizations.
The following website has been deployed and you can use it over here - https://prism-gi6z.onrender.com/ Make sure to enable 'Open Source' model option as Claude API connection has been revoked

Quick Start (One Command)

docker run -p 8000:8000 \
  -e ANTHROPIC_API_KEY=your_anthropic_key \
  talk-to-data

Then open http://localhost:8000

---

Features

Feature	Description
Natural Language Queries	Ask questions in plain English, get SQL results
Multi-Model AI	Claude (paid) or Groq (free) - switch in settings
Universal Database	SQLite, PostgreSQL, MySQL - one config change
Self-Correcting	Automatically retries failed queries with fixes
Deep Analysis	Auto-generates charts and insights from results
Read-Only Safety	Blocks all write operations to protect your data
Mobile Optimized	Responsive design works on all devices
Single Container	Frontend + Backend + Demo DB in one image

---

Configuration

Environment Variables

Variable	Required	Description
ANTHROPIC_API_KEY	Yes	Your Anthropic API key for Claude models
GROQ_API_KEY	No	Groq API key for free tier models
DATABASE_URL	No	External database connection string
PORT	No	Server port (default: 8000)

Connect Your Database

PostgreSQL:

docker run -p 8000:8000 \
  -e ANTHROPIC_API_KEY=your_key \
  -e DATABASE_URL=postgresql://user:pass@host:5432/dbname \
  talk-to-data

MySQL:

docker run -p 8000:8000 \
  -e ANTHROPIC_API_KEY=your_key \
  -e DATABASE_URL=mysql://user:pass@host:3306/dbname \
  talk-to-data

SQLite (default): No DATABASE_URL needed - uses the built-in demo database.

---

Cloud Deployment

Railway

1. Push your code to GitHub
2. Go to railway.app and connect your repo
3. Railway auto-detects the railway.json configuration
4. Add environment variables in the dashboard:
   • ANTHROPIC_API_KEY
   • GROQ_API_KEY (optional)
   • DATABASE_URL (optional)
5. Deploy

Render

1. Push your code to GitHub
2. Go to render.com and create a new Web Service
3. Connect your repo - Render auto-detects render.yaml
4. Add environment variables in the dashboard
5. Deploy

Any Docker Host

# Build
docker build -t talk-to-data .

# Run
docker run -d -p 8000:8000 \
  --name talk-to-data \
  -e ANTHROPIC_API_KEY=your_key \
  -e DATABASE_URL=your_db_url \
  talk-to-data

---

Local Development

Prerequisites

• Python 3.11+
• Node.js 20+

Setup

# Clone the repo
git clone https://github.com/your-repo/talk-to-data.git
cd talk-to-data

# Create .env file
echo "ANTHROPIC_API_KEY=your_key" > .env

# Install backend
cd backend
pip install -r ../requirements.txt

# Install frontend
cd ../frontend
npm install

# Run backend (terminal 1)
cd backend && python main.py

# Run frontend (terminal 2)
cd frontend && npm run dev

---

API Endpoints

Endpoint	Method	Description
/api/health	GET	Health check with database ping
/api/query	POST	Execute a natural language query
/api/query/stream	POST	Stream query results (SSE)
/api/analyze	POST	Generate chart from query results

---

Architecture

talk-to-data/
├── backend/
│   ├── main.py           # FastAPI server
│   ├── config.py         # Centralized settings
│   ├── agent_engine.py   # Query orchestration
│   ├── llm_client.py     # AI model interface
│   ├── database_utils.py # Universal DB connector
│   └── analysis_engine.py# Chart generation
├── frontend/
│   └── src/              # React application
├── assets/
│   └── chinook.db        # Demo database
├── Dockerfile            # Multi-stage build
├── railway.json          # Railway config
└── render.yaml           # Render config

---

License

MIT