seekdb is an embedded, MySQL-compatible database that combines vector, full-text, and scalar data in one engine. It targets AI agent workloads with features like copy-on-write sandboxes and hybrid search, and provides Python SDKs and SQL examples.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
seekdb is a MySQL-compatible, embedded or server-capable database that unifies vector, full-text, and scalar data in one engine. It supports hybrid search and is designed for agent storage and AI workloads. It emphasizes Copy-on-Write sandboxes and streaming write + concurrent search for agent-style workloads. The project is licensed under Apache-2.0 and is written in C++.
How it works
The repository describes an architecture with an async index pipeline for incremental vector index updates and a two-level HNSW for search. It supports a Change Stream to decouple DML from index construction, allowing writes to proceed without waiting for index builds. It combines vector indexes, full-text indexes, and scalar filters within a single SQL execution plan. It can run in embedded mode in-process or as a server, and also integrates with MySQL-compatible workflows.
Getting started
The README provides multiple installation paths:
- Python (Recommended for AI/ML)
pip install -U pyseekdb - Cloud (Zero Install)
curl -X POST https://d0.seekdb.ai/api/v1/instances - Docker (Quick Testing)
docker run -d \ --name seekdb \ -p 2881:2881 \ -p 2886:2886 \ -v ./data:/var/lib/oceanbase \ oceanbase/seekdb:latest - Binary (Standalone)
# Linux curl -fsSL https://obportal.s3.ap-southeast-1.amazonaws.com/download-center/opensource/seekdb/seekdb_install.sh | bash # macOS brew tap oceanbase/seekdb brew install seekdb
It also mentions a 30-second demo and provides code examples for Python SDK usage and SQL usage demonstrating a hybrid query with a vector column and a full-text index.
Recent releases
Latest releases listed in RELEASes:
- v1.3.0 (2026-05-25): major release for multi-platform coverage; RPM: seekdb-1.3.0.0-100000092026051510
- v1.2.0 (2026-04-15): major milestone; RPM: seekdb-1.2.0.0-100000232026041319
- v1.1.0 (2026-01-30): macOS build support; RPM: seekdb-1.1.0.0-100000142026013001
- v1.0.1 (2025-12-29): vector index enhancements; RPM: seekdb-1.0.1.0-100000392025122619
- v1.0.0 (2025-11-13): release notes; RPM: seekdb-1.0.0.0-100000262025111218
Traction
The repository shows 2856 stars, 309 forks, and 452 open issues.
Behind the repo
Not applicable per facts: no linked startup or company profile is provided in the given data.
Caveats
License is Apache-2.0. Creation date is 2025-10-21 and last push is 2026-08-04. The README references multiple build and deployment paths, but no detailed licensing caveats beyond the standard Apache-2.0 note is provided in the supplied facts.