Access and Query UniProt & EBI Protein Data
This MCP server unifies UniProtKB and EBI Proteins API access, offering search, BLAST, and proteomics tools with SQLite caching.
Why it matters
Gain unified access to comprehensive protein sequence, functional, and structural data from UniProtKB and EBI Proteins API. This asset enables advanced querying, bulk loading, and analysis of protein information with intelligent caching.
Outcomes
What it gets done
Perform advanced searches and retrieve individual protein entries.
Bulk load and cache large protein datasets for efficient querying.
Map protein IDs across different databases and perform BLAST searches.
Analyze cached protein data using SQL and access detailed annotations.
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/vb-uniprot | bash Capabilities
Tools your agent gets
Advanced search across UniProtKB with comprehensive filtering and pagination
Bulk loading tool for large datasets with automatic caching
Retrieve individual UniProtKB entries by accession number
Map IDs across different database systems
Perform BLAST searches against UniProtKB
Detailed protein information from EBI Proteins API
Protein sequence features and annotations
Protein sequence variations and disease variants
Overview
UniProt MCP Server
This MCP server gives unified, SQLite-cached access to UniProtKB and the EBI Proteins API through eleven tools covering search, BLAST, ID mapping, and protein feature, variation, proteomics, and genome data. Use it when a workflow needs protein sequence, annotation, or structural data from UniProt or EBI and benefits from cached, SQL-queryable results.
What it does
This server provides unified MCP access to both UniProtKB and the EBI Proteins API, covering protein sequence data, functional annotations, and structural information. It layers advanced SQLite-backed caching on top of both APIs so repeated or complex queries do not repeatedly hit the upstream services, and it normalizes results across data formats (JSON, TSV, FASTA, XML) into a consistent interface for an AI assistant.
When to use - and when NOT to
Use this when a workflow needs protein-level bioinformatics data - searching UniProtKB by organism or review status, mapping IDs across database systems, running BLAST searches, or pulling variation, proteomics, or genome-coordinate data from EBI - and you want that data cached and queryable via SQL rather than re-fetched on every request. It is not a general biology knowledge tool; it is scoped specifically to UniProt and EBI Proteins API data.
Inputs and outputs
Eleven tools are exposed: uniprot_search for advanced filtered search with pagination, uniprot_stream for bulk loading large datasets with automatic caching, uniprot_entry for retrieving a specific entry by accession, uniprot_id_mapping for cross-database ID mapping, uniprot_blast for BLAST searches against UniProtKB, proteins_api_details/proteins_api_features/proteins_api_variation/proteins_api_proteomics/proteins_api_genome for detailed protein information, sequence features, disease variants, proteomics studies, and genome coordinate mappings respectively, and data_manager for querying and managing the cached datasets directly.
Integrations
Built on Cloudflare Workers with Durable Objects for scalability, exposing endpoints at /mcp and /sse; no API key is required since both UniProt and EBI Proteins API are open access. It is used via mcp-remote pointed at a deployed Worker URL, or run locally with npm run dev and deployed with npm run deploy.
Who it's for
Bioinformaticians, computational biology researchers, and developers building protein-data tooling who want a single, cached, SQL-queryable interface over UniProt and EBI Proteins API rather than juggling both APIs directly. The data_manager tool in particular suits workflows that stage many proteins with uniprot_stream and then need to run further SQL analysis across that cached dataset without re-querying the upstream APIs each time.
npx mcp-remote http://localhost:8787/sse
Source README
A comprehensive MCP server providing unified access to UniProtKB and EBI Proteins API for working with protein sequence data, functional annotations, and structural information with advanced caching capabilities.
Installation
NPM Install
npm install
Development
npm run dev
Deploy to Cloudflare
npm run deploy
MCP Remote Proxy
npx mcp-remote http://localhost:8787/sse
Configuration
Claude Desktop
{
"mcpServers": {
"uniprot": {
"command": "npx",
"args": [
"mcp-remote",
"http://localhost:8787/sse"
]
}
}
}
Available Tools
| Tool | Description |
|---|---|
uniprot_search |
Advanced search across UniProtKB with comprehensive filtering and pagination |
uniprot_stream |
Bulk loading tool for large datasets with automatic caching |
uniprot_entry |
Retrieve individual UniProtKB entries by accession number |
uniprot_id_mapping |
Map IDs across different database systems |
uniprot_blast |
Perform BLAST searches against UniProtKB |
proteins_api_details |
Detailed protein information from EBI Proteins API |
proteins_api_features |
Protein sequence features and annotations |
proteins_api_variation |
Protein sequence variations and disease variants |
proteins_api_proteomics |
Proteomics data from various studies |
proteins_api_genome |
Genome coordinate mappings |
data_manager |
Query, analyze, and manage cached datasets |
Features
- Unified interface for UniProt and EBI Proteins API
- Advanced caching with SQLite for complex queries
- Smart query generation with automatic suggestions
- Intelligent pagination for small datasets
- Scalable architecture based on Cloudflare Workers
- API rate limiting handling
- Support for multiple data formats (JSON, TSV, FASTA, XML)
- Automatic table creation and data normalization
- Bulk ID mapping across database systems
- BLAST search capabilities
Usage Examples
Search for human proteins: organism_id:9606 AND reviewed:true
Get protein details for accession P04637
Stage and query multiple proteins with SQL
Find cancer-related proteins and analyze with keywords
Search protein kinase family members
Notes
Built on Cloudflare Workers with Durable Objects for scalability. No API keys required as both UniProt and EBI Proteins API are open access. Server endpoints are available at /mcp and /sse. Includes comprehensive SQL query capabilities for analyzing cached data.
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.