Morphik MCP
Enables interaction with the Morphik multi-modal database system for document ingestion, retrieval, querying, and management. Supports text and file ingestion, semantic search with LLM-powered completions, and file system navigation with security controls.
README
Morphik MCP
A Model Context Protocol (MCP) server implementation for Morphik multi-modal database.
Overview
This MCP server allows Claude and other MCP-compatible AI assistants to interact with the Morphik database system, enabling:
- Document ingestion (text and files)
- Document retrieval (by relevance to queries)
- Document querying with LLM-powered completions
- Document management (listing, getting, deleting)
- File system navigation and file ingestion from paths
The server supports two transport modes:
- stdio (default): Standard input/output for use with Claude Desktop and other MCP clients
- Streamable HTTPS: HTTP/HTTPS endpoint for web-based integrations
Prerequisites
You need access to either:
- A local Morphik server running on localhost:8000, or
- The Morphik cloud platform
Installation & Usage
Option 1: Run directly with npx (recommended)
# Connect to local Morphik server
npx morphik-mcp
# Connect to Morphik cloud platform (replace with your actual URI)
npx morphik-mcp --uri=https://api.morphik.ai
# Specify allowed directories for file operations (comma-separated)
npx morphik-mcp --allowed-dir=~/Documents,~/Downloads
Option 2: Global installation
npm install -g morphik-mcp
# Connect to local Morphik server
morphik-mcp
# Connect to Morphik cloud platform
morphik-mcp --uri=https://api.morphik.ai
# Specify allowed directories for file operations
morphik-mcp --allowed-dir=~/Documents,~/Downloads
Option 3: Local development
# Clone the repository
git clone https://github.com/morphik-org/morphik-npm-mcp.git
cd morphik-npm-mcp
# Install dependencies
npm install
# Build the project
npm run build
# Start the server (local Morphik)
npm start
# Start the server with file operations enabled
node build/index.js --allowed-dir=~/Documents,~/Downloads
The server runs on standard input/output streams by default and can be used with MCP clients like Claude.
For HTTP/HTTPS mode (streamable), see the streamable_https.md documentation.
Usage with Claude Desktop
Add this to your Claude Desktop configuration file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"morphik-local": {
"command": "npx",
"args": ["-y", "morphik-mcp"]
},
"morphik-cloud": {
"command": "npx",
"args": ["-y", "morphik-mcp", "--uri=https://api.morphik.ai"]
},
"morphik-with-files": {
"command": "npx",
"args": ["-y", "morphik-mcp", "--allowed-dir=~/Documents,~/Downloads"]
}
}
}
MCP Tools
See MCP_TOOLS.md for detailed parameter docs and response formats.
The server currently provides the following tools, with additional capabilities on the way:
1. Document Ingestion
-
ingest-text: Ingest a text document into Morphik- Parameters: content, filename (optional), metadata (optional), apiKey
-
ingest-file-from-path: Ingest a file from the server's filesystem into Morphik- Parameters: path, metadata (optional), rules (optional), folderName (optional), endUserId (optional), useColpali (optional)
-
ingest-file-from-base64: Upload a file by providing its filename plus base64-encoded bytes (works in HTTP transports where disk access isn't possible)- Parameters: filename, base64Content, metadata (optional), rules (optional), folderName (optional), endUserId (optional), useColpali (optional)
-
ingest-files-from-paths: Batch ingest multiple files from the server's filesystem- Parameters: paths, metadata (optional), rules (optional), folderName (optional), endUserId (optional), useColpali (optional)
2. Document Retrieval & Navigation
-
retrieve-chunks: Retrieve the most relevant "pages" (text or image chunks) for a query, identical to thefind_relevant_pagestool in@morphik-app- Parameters: query, filters (optional), k (optional), minScore (optional), folderName (optional), endUserId (optional), useColpali/useReranking (optional), padding (optional for adjacent context)
-
retrieve-docs: Retrieve relevant documents from Morphik based on a query- Parameters: query, filters (optional), k (optional), minScore (optional), folderName (optional), endUserId (optional)
-
search-documents: Search for documents by filename or title using full-text search- Parameters: query, limit (optional, 1-100), folderName (optional), endUserId (optional)
-
get-pages-in-range: Fetch an inclusive range of pages (max 10) from a document, matchingget_all_pages_within_range_from_documentin@morphik-app- Parameters: documentId, startPage, endPage, folderName (optional), endUserId (optional)
3. Document Management & Metadata
-
get-document: Get a specific document from Morphik by ID- Parameters: documentId, apiKey
-
delete-document: Delete a document from Morphik by ID- Parameters: documentId, apiKey
-
check-ingestion-status: Check the processing status for an ingested document- Parameters: documentId, apiKey
-
morphik-filters: View, set, or clear typed metadata filters that automatically apply to retrieval/list operations (supportseq,regex,number_range,date_range)- Parameters: action, expression/rules payload, logic (optional)
-
list-documents: Updated tool using/documents/list_docswithinclude_total_count,return_documents,fields, and folder/end-user scoping- Parameters: skip, limit, filters/metadataFilters, get_count, folderName(s), endUserId, sortBy/sortDirection, fields
4. File System Navigation
-
list-allowed-directories: List directories that the server is allowed to access- Parameters: none
-
list-directory: List files and subdirectories in a specific directory- Parameters: path
-
search-files: Search for files matching a pattern in a directory and its subdirectories- Parameters: path, pattern, excludePatterns (optional)
-
get-file-info: Get detailed information about a file or directory- Parameters: path
File Operations Security
For security reasons, file operations are restricted to directories explicitly allowed when starting the server using the --allowed-dir parameter. If no directories are specified, only the user's home directory will be accessible.
The server validates all file paths to ensure they're within allowed directories, preventing access to sensitive system files. Symlinks are also checked to ensure they don't point outside allowed directories.
Development
Build the project:
npm run build
Run tests:
npm test
Run manual tests:
npm run test:manual
License
ISC
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.