Enrichr MCP Server

The Enrichr MCP Server is a community-built Model Context Protocol server that wraps the public Enrichr API from the Ma'ayan Lab. It lets LLMs perform gene set enrichment analysis against more than 200 curated gene set libraries spanning Gene Ontology (Biological Process, Molecular Function, Cellular Component), pathway databases (KEGG, Reactome, WikiPathways, MSigDB Hallmark), transcription factor targets (ChEA), disease and phenotype catalogs (GWAS Catalog, Human Phenotype Ontology), drug perturbation libraries (DrugBank, L1000), and cell-type/tissue references (CellMarker, ARCHS4, GTEx).

The server exposes two tools: suggest_libraries for discovering relevant libraries from a natural-language research question, and enrichr_analysis for running parallel enrichment across multiple libraries and returning only statistically significant results (corrected p-value < 0.05). Output can be returned in detailed, compact, or minimal formats to manage LLM token budgets, and results can be exported to TSV for downstream use.

This is a third-party project (not an official Ma'ayan Lab release) written in TypeScript and distributed via npm. It requires no API key because Enrichr is a free public service, but users should cite the original Enrichr publications when publishing results derived from the tool.

Tools

Prerequisites

• Node.js 18 or newer
• Internet connection (the server calls the public Enrichr API)
• No API key or account required

Install via Claude Code

claude mcp add enrichr-mcp-server -- npx -y enrichr-mcp-server

Manual configuration (Claude Desktop, Cursor, VS Code)

Add the following to your MCP client config:

{
  "mcpServers": {
    "enrichr-server": {
      "command": "npx",
      "args": ["-y", "enrichr-mcp-server"]
    }
  }
}

Optional environment variables

• ENRICHR_LIBRARIES: comma-separated list of default libraries to query
• ENRICHR_MAX_TERMS: maximum terms returned per library
• ENRICHR_FORMAT: output format (detailed, compact, or minimal)
• ENRICHR_OUTPUT_FILE: path for TSV export of results

Claude Desktop bundle

A prebuilt .mcpb bundle is available from the GitHub releases page and can be installed directly from Claude Desktop settings.

Default library set

If no libraries are specified, the server queries 10 commonly used libraries including GO_Biological_Process_2025, KEGG_2021_Human, Reactome_2022, MSigDB_Hallmark_2020, ChEA_2022, GWAS_Catalog_2023, Human_Phenotype_Ontology, STRING_Interactions_2023, DrugBank_2022, and CellMarker_2024.

• Interpret differentially expressed gene lists from RNA-seq or microarray experiments by identifying overrepresented GO Biological Processes and pathways
• Discover candidate upstream transcription factors for a gene signature using ChEA and ENCODE libraries
• Map a gene list to associated diseases and traits via GWAS Catalog and Human Phenotype Ontology
• Identify drugs or chemical perturbations that produce similar (or opposing) expression signatures via DrugBank and L1000 libraries
• Annotate single-cell cluster marker lists with likely cell types using CellMarker and ARCHS4 tissue libraries

• "Run Enrichr on these 50 upregulated genes from my RNA-seq experiment against KEGG and Reactome and summarize the top significant pathways."
• "Suggest the most relevant Enrichr libraries for studying neurodegeneration in human brain tissue."
• "Take this gene list and find candidate transcription factors that may regulate them using ChEA."
• "Run GO Biological Process enrichment on BRCA1, BRCA2, TP53, ATM, CHEK2, PALB2 and export the results to a TSV file."
• "Which drugs in DrugBank or L1000 produce expression signatures similar to this list of inflammation genes?"

• Wraps a well-established, widely cited bioinformatics service (Enrichr) with 200+ curated libraries
• No API key or paid plan required; uses the free public Enrichr API
• Parallel multi-library querying with three output formats to manage LLM token usage
• Built-in library suggestion tool helps the LLM pick appropriate databases for a research question

• Community-maintained third-party project, not an official Ma'ayan Lab release
• Small project (single maintainer, low star count) so long-term support is uncertain
• Depends on the public Enrichr API; rate limits or outages will directly impact functionality

• BioMCP: broader biomedical MCP server covering literature, variants, and clinical trials
• Direct use of the Enrichr REST API from a custom MCP tool or script
• gseapy: Python library for gene set enrichment that can be wrapped in a custom MCP server