Repository: databendlabs/databend

Stars: 9254

README.md Copy Chapter

<h1 align="center">Databend</h1>
<h3 align="center">Enterprise Data Warehouse for AI Agents</h3>
<p align="center">Large-scale analytics, vector search, full-text search — with flexible agent orchestration and secure Python UDF sandboxes. Built for enterprise AI workloads.</p>

<div align="center">

<a href="https://databend.com/">☁️ Try Cloud</a> •
<a href="#-quick-start">🚀 Quick Start</a> •
<a href="https://docs.databend.com/">📖 Documentation</a> •
<a href="https://link.databend.com/join-slack">💬 Slack</a>

<br><br>

<a href="https://github.com/databendlabs/databend/actions/workflows/release.yml">
<img src="https://img.shields.io/github/actions/workflow/status/datafuselabs/databend/release.yml?branch=main" alt="CI Status" />
</a>
<img src="https://img.shields.io/badge/Platform-Linux%2C%20macOS%2C%20ARM-green.svg?style=flat" alt="Platform" />

</div>

<br>

<img src="https://github.com/user-attachments/assets/4c288d5c-9365-44f7-8cde-b2c7ebe15622" alt="databend" width="100%" />

💡 Why Databend? Copy Chapter

Databend is an open-source enterprise data warehouse built in Rust.

Core capabilities: Analytics, vector search, full-text search, auto schema evolution — unified in one engine.

Agent-ready: Sandbox UDFs for agent logic, SQL for orchestration, transactions for reliability, branching for safe experimentation on production data.

| | |
| :--- | :--- |
| 📊 Core Engine<br>Analytics, vector search, full-text search, auto schema evolution, transactions. | 🤖 Agent-Ready<br>Sandbox UDF + SQL orchestration. Build and run agents on your enterprise data. |
| 🏢 Enterprise Scale<br>Elastic compute, cloud native. S3/Azure/GCS. | 🌿 Branching<br>Git-like data versioning. Agents safely operate on production snapshots. |

!Databend Architecture

⚡ Quick Start Copy Chapter

1. Cloud (Recommended) Copy Section

2. Local (Python) Copy Section

pip install databend

import databend
ctx = databend.SessionContext()
ctx.sql("SELECT 'Hello, Databend!'").show()

3. Docker Copy Section

docker run -p 8000:8000 datafuselabs/databend

🤖 Agent-Ready Architecture Copy Chapter

Databend's Sandbox UDF enables flexible agent orchestration with a three-layer architecture:

- Control Plane: Resource scheduling, permission validation, sandbox lifecycle management
- Execution Plane (Databend): SQL orchestration, issues requests via Arrow Flight
- Compute Plane (Sandbox Workers): Isolated sandboxes running your agent logic

-- Define your agent logic
CREATE FUNCTION my_agent(input STRING) RETURNS STRING
LANGUAGE python HANDLER = 'run'
AS $$
def run(input):
    # Your agent logic: LLM calls, tool use, reasoning...
    return response
$$;
-- Orchestrate agents with SQL
SELECT my_agent(question) FROM tasks;

🚀 Use Cases Copy Chapter

- AI Agents: Sandbox UDF + SQL orchestration + branching for safe operations
- Analytics & BI: Large-scale SQL analytics — Learn more
- Search & RAG: Vector + full-text search — Learn more

🤝 Community & Support Copy Chapter

- 📖 Documentation
- 💬 Join Slack
- 🐛 Issue Tracker
- 🗺️ Roadmap

Contributors are immortalized in the system.contributors table 🏆

📄 License Copy Chapter

Apache 2.0 + Elastic 2.0 | Licensing FAQ

---

<div align="center">
<strong>Enterprise warehouse, agent ready</strong><br>
<a href="https://databend.com">🌐 Website</a> •
<a href="https://x.com/DatabendLabs">🐦 Twitter</a>
</div>