BenderV/ada

Accelerate your data analysis with AI.

ADA: Accelerate Data Analysis, with AI

Ada is the best AI tool to explore your data in a safe, fast & accessible way.

Features

• Easy set up: Run Ada locally in 5 minutes
• Human Friendly: Query in English, Ada will take care of the rest.
• “Take Over” Mode: Built-in Editor to run SQL
• Built-in Privacy & Safety: Use with a peace of mind
• DBT Support: Upload DBT catalog and manifest to leverage DBT (experimental)

Demo

Check out this 2 min demo video to see Ada in action. Or better 👉 try it yourself

Quick Start Guide

Pre-requisites

ℹ️ Note: get an OpenAI API key. If you don't have one, get it here.

Docker installation

docker-compose up -d

Manual installation

Backend Setup (in /service directory)

1. Install Dependencies:

   poetry install

2. Set Environment Variables: in service/.env.sh file:

   export AUTOCHAT_MODEL=<Your_OpenAI_Model>
   export OPENAI_API_KEY=<Your_OpenAI_API_Key>
   export DATABASE_URL=<Your_Postgres_Database_URL>

3. Run the Backend:

   bash run.sh

Frontend Setup (in /view directory)

Install Dependencies

yarn

Run the front-end

yarn dev

After completing the steps, open your browser and visit: http://localhost:5173

Privacy & Safety

We have developed built-in features to protect privacy and insure safety

• Preview only: Model only have preview of the requested data
• “Privacy protection”: hide fields about name, location
• “Safe mode”: limit to read-only query

Tech Stacks

• AI Interaction: Using Autochat library with OpenAI GPT-4 API.
• Frontend: Built using Vue3 and Vite
• Backend: Developed in Python
• Database: Postgres

FAQs

• Q: How secure is my data with Ada / OpenAI?
• A: Since it's open source, you can run Ada on your own server and keep your data private. OpenAI's API is also secure and encrypted, and they don't use your data for training with the API.

Troubleshooting & Contribution

For any issues, please open an issue on GitHub or contact me on Twitter.

License

LICENSE - AGPL