Databend

Unified Multimodal Database for Any Data at Any Scale.

A next-generation cloud-native warehouse built in Rust. Open-source, Snowflake-compatible, and unifying BI, AI, Search, Geo, and Stream.

☁️ Try Cloud • 🚀 Quick Start • 📖 Documentation • 💬 Slack

💡 Why Databend?

Databend is an open-source unified multimodal database built in Rust. It empowers Analytics, AI, Search, and Geo workloads on a single platform directly from object storage.

• Unified Engine: One optimizer and runtime for all data types (Structured, Semi-structured, Vector).
• Native Pipelines: Built-in Stream and Task for automated data cleaning and transformation.
• Cloud Native: Stateless compute nodes over object storage (S3, GCS, Azure) with full ACID support.
• High Performance: Vectorized execution and Zero-Copy processing.

⚡ Quick Start

1. Cloud (Recommended)

Start for free on Databend Cloud - Production-ready in 60 seconds.

2. Local (Python)

Ideal for development and testing:

pip install databend

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

3. Docker

Run the full warehouse locally:

docker run -p 8000:8000 datafuselabs/databend

🚀 Use Cases

• BI & Analytics: High-speed SQL on massive datasets. See Query Processing.
• AI & Vectors: Built-in vector search and embedding management. See Vector Database.
• Full-Text Search: Fast indexing and retrieval on text and semi-structured data (JSON). See JSON Search.
• Geospatial: Advanced geo-analytics and mapping. See Geospatial Analysis.
• Stream & Task: Continuous data ingestion and transformation. See Real-Time ETL.

🤝 Community & Support

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

Contributors are immortalized in the system.contributors table! 🏆

📄 License

Apache 2.0 + Elastic 2.0 | Licensing FAQ

---

Built by engineers who redefine what's possible with data
🌐 Website • 🐦 Twitter