MarkLogic MCP and Retrieval API

The MarkLogic MCP project enables AI assistants (like GitHub Copilot, AWS Quick, Claude Desktop, and others) to search and retrieve documents from MarkLogic databases using the Model Context Protocol.

Key Components:

MarkLogic Retrieval API (MRA): REST API for multi-modal search and document retrieval
MCP Server: Hyperforge-based server that exposes MRA as MCP tools
MCP Deployer: Tools for deploying the Retrieval API to MarkLogic

Features

Semantic Search — Vector and full-text search with configurable relevance
Label-based Filtering — Organize and filter documents by corpus labels
Document Retrieval — Fetch full documents by URI with metadata
Multiple Authentication Methods — Basic, Digest, OAuth/Keycloak support
Docker Ready — Easy deployment with Docker Compose
AI Assistant Integration — Works with any MCP-compatible client

---

Getting Started:

See the README in this repository for information on installation and configuration.