Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@STRING MCP Server Find the interaction partners for the human protein TP53"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Model Context Protocol (MCP) server providing access to the STRING protein-protein interaction database.
9 tools for protein interaction analysis:
Mapping protein identifiers to STRING IDs
Retrieving protein interaction networks
Finding interaction partners
Functional enrichment analysis (GO, KEGG, pathways)
PPI enrichment testing
Cross-species homology mapping
Protein name resolution
Full endpoint documentation available in ENDPOINTS.md.
Related MCP server: UniProt MCP Server
npm install git+https://github.com/munch-group/string-mcp.git
cd node_modules/string-mcp
npm run buildgit clone https://github.com/munch-group/string-mcp.git
cd string-mcp
npm install
npm run build# Using built version
claude mcp add string node /path/to/string-mcp/build/index.js
# Or from npm global install
claude mcp add string npx -y string-mcpAdd to your Claude Desktop config:
{
"mcpServers": {
"string": {
"command": "node",
"args": ["/path/to/string-mcp/build/index.js"]
}
}
}Human: 9606
Mouse: 10090
Rat: 10116
Fruit fly: 7227
C. elegans: 6239
Yeast: 4932
npm run build@mcp/rest-utils - Shared REST utilities
@modelcontextprotocol/sdk - MCP SDK
zod - Schema validation
ISC
Perform functional enrichment analysis for a set of proteins. Tests for over-representation in Gene Ontology terms, KEGG pathways, etc.
Inputs: identifiers (required), species, background.
Get homology information for proteins across species using STRING.
Inputs: identifiers (required), species, target_species.
Get the best homology match for proteins in a target species.
Inputs: identifiers (required), target_species (required), species.
Get all STRING interaction partners for your proteins. Returns a list of interacting proteins with confidence scores.
Inputs: identifiers (required), species, limit, required_score, +1 more.
Retrieve protein-protein interaction network for given proteins. Returns network edges with confidence scores.
Inputs: identifiers (required), species, required_score, network_type, +1 more.
Test if your protein set has more interactions than expected by chance. Returns enrichment p-value.
Inputs: identifiers (required), species, required_score.
Map protein names, synonyms, and UniProt identifiers to STRING identifiers. Species parameter uses NCBI taxon IDs (e.g., 9606 for human, 10090 for mouse).
Inputs: identifiers (required), species, limit, echo_query.
Get the current version of the STRING database.
Inputs: no named parameters.
Resolve protein names to their preferred names in STRING database.
Inputs: identifiers (required), species.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 9 tools
Most tools have distinct purposes, such as get_enrichment for functional analysis and get_network for interaction networks. However, get_homology and get_homology_best could be confused as they both handle homology, with the latter being a more specific version, creating minor ambiguity.
All tool names follow a consistent verb_noun pattern starting with 'get_' or 'resolve_', such as get_enrichment and resolve_proteins. This uniformity makes the set predictable and easy to navigate.
With 9 tools, the server is well-scoped for protein analysis tasks, covering key operations like enrichment, homology, interactions, and ID mapping. Each tool serves a clear purpose without being overwhelming.
The tool set covers essential protein analysis workflows, including enrichment, homology, interactions, and ID resolution. A minor gap exists in lacking update or delete operations, but this is reasonable for a read-focused database query server.
Query STRING interactions, enrichment, annotations, homology, and PPI networks.
STRING-DB MCP — protein-protein interaction networks.
Link compounds to protein targets, rank bioactivity, and look up drug mechanisms and indications.
Protein research over UniProtKB: search by function, fetch curated records, map IDs, proteomes.
