Jooq MCP

What jooq-mcp MCP server does

The jooq-mcp MCP server connects MCP-compatible AI clients to documentation about jOOQ, a Java library for building SQL queries and working with databases. Its tools support targeted questions about query construction, supported databases, code generation, Query DSL statements, SQL operations, and advanced jOOQ features.

It is intended for development workflows where an assistant needs to look up jOOQ-specific behavior or examples. Typical questions include how to build a SELECT or INSERT statement, configure code generation, use transactions, or determine database support.

How it works

The application is implemented as a Spring Boot service using Spring AI's MCP server support. Tool methods are registered from the documentation service and exposed to connected MCP clients. The server uses synchronous tool handling.

Documentation can be stored locally under src/main/resources/docs/. A crawler fetches documentation, while a fetcher parses and extracts usable content. Local search uses an in-memory inverted index with TF-IDF relevance scoring. Code examples are extracted and categorized so requests for query-building examples can return focused snippets rather than only broad pages.

Caffeine caching is used to reduce repeated work. The repository also includes crawler limits for depth, URLs per section, request timeout, and cache duration. These settings allow the documentation collection process to be bounded and reused.

Setup and configuration

Running the project requires Java 21 or later and Maven 3.6 or later. The documented local workflow is:

1. Build with ./mvnw clean install.
2. Start the application with ./mvnw spring-boot:run.
3. Connect an MCP-compatible client to the running service.

The application can also be packaged into a JAR and run in Docker on port 8080. Fly.io deployment is supported by the repository configuration, and the application exposes /actuator/health for health checks.

Configuration is provided through application.properties. Supported settings include the MCP server name, version, synchronous mode, tool capability, Caffeine cache policy, crawler limits, and Tomcat response-buffer limits. The sample cache is configured for up to 100 entries with a one-hour expiration, while crawled documentation is configured with a 24-hour cache duration.

Tools and capabilities

The jooq-mcp MCP server exposes these documented tools:

• searchDocumentation searches jOOQ documentation by topic or keyword.
• getSqlExamples retrieves examples for specified SQL operations.
• getCodeGenerationGuide returns guidance for jOOQ code generation.
• getDatabaseSupport provides database-specific information.
• getQueryDslReference looks up Query DSL material for statement types.
• getAdvancedFeatures retrieves documentation about advanced jOOQ functionality.

Together, these tools cover reference lookup, example discovery, configuration research, and feature-specific documentation queries.

Limitations and notes

The provided material describes documentation access rather than database execution or schema inspection. The server retrieves and searches jOOQ reference content; it does not claim to run SQL, connect to a user database, or modify a project.

Documentation quality and coverage depend on the locally stored or crawled source material. The repository includes local indexing, crawling, parsing, and caching components, but the excerpt does not specify an automatic refresh schedule beyond the configurable cache duration. Client setup details are not provided for any particular MCP application, so connection steps depend on the client being used.