This organization holds code samples, quickstarts and demo applications for building with MariaDB.
Many applications run one system for transactions, another for analytics and a third for vectors, with pipelines to keep the data in sync. AI agents make this harder, because they need current data from all of those systems at once.
MariaDB keeps these workloads in one open source database, so your application, your reports and your agents query the same live data with the same SQL:
- Transactions with row storage in the InnoDB storage engine.
- Analytics with columnar storage in MariaDB ColumnStore.
- Vector search with a native
VECTORdata type and index, available in MariaDB Server since 11.7 (vector overview). - JSON documents with the built-in JSON functions.
- AI tools such as the MariaDB MCP server, which lets AI assistants query MariaDB through the Model Context Protocol, and vector store integrations for LangChain, LlamaIndex and Spring AI.
MariaDB Server is open source under the GPL v2 license and runs where you choose: on your laptop, on your own servers or in any cloud.
- Read the MariaDB documentation for guides, SQL reference and connectors.
- Run MariaDB locally with the official Docker image, or download MariaDB Server.
- Clone a quickstart: Java, Python, Node.js, PHP or TypeScript.
- Try vector search with LangChain in the vector search workshop.
Open issues and pull requests directly in each repository. If you built a project with MariaDB that other developers can learn from, send a short description and a link to developers@mariadb.com.
