metazense/dazense repository overview

⁠dazense - The #1 Open-Source Analytics Agent

dazense is a framework to build and deploy analytics agents. Create context for your analytics agent with the dazense-core CLI, then deploy a chat UI for anyone to interact with your data.

🌐 Website⁠ · 📚 Documentation⁠ · 💬 Slack⁠ · 🐙 GitHub⁠

⁠Docker

⁠Supported Tags

• latest - Latest stable release
• commit-hash - Specific commit hash tags

⁠Supported Architectures

• linux/amd64

⁠Base Image

• Python 3.12 (slim) with Node.js 24 and Bun

⁠Quick Start

⁠Using Docker Run (SQLite database)

docker run -d \
  --name dazense \
  -p 5005:5005 \
  -e OPENAI_API_KEY=sk-... \
  -v /path/to/your/project:/app/project \
  metazense/dazense:latest

⁠Environment Variables

* At least one LLM API key is required to make AI queries.

⁠Ports

⁠Volumes

Mount your dazense project directory to make it available to the agent:

-v /path/to/your/dazense-project:/app/project

Then set DAZENSE_DEFAULT_PROJECT_PATH=/app/project, you can also use the example project by setting DAZENSE_DEFAULT_PROJECT_PATH=/app/example.

⁠Docker run example

docker run -d \
  --name dazense \
  -p 5005:5005 \
  -e DAZENSE_DEFAULT_PROJECT_PATH=/app/project \
  metazense/dazense:latest

Then navigate to http://localhost:5005⁠ to access the UI (or to any URL you configured).

⁠Key Features

• 🧱 Open Context Builder — Create file-system context for your agent
• 🏳️ Data Stack Agnostic — Works with any data warehouse, stack, LLM
• 🔒 Self-hosted & Secure — Use your own LLM keys for maximum data security
• 🤖 Natural Language to Insights — Ask questions in plain English
• 📊 Native Data Visualization — Create visualizations directly in chat

⁠License

Apache 2.0 - See LICENSE⁠