Ensembl Mcp Server
Give Claude Code or Cursor live Ensembl REST access for gene lookup, sequences, variant effect prediction, and ortholog mapping while you prototype bioinformatics or health-research tools.
Overview
io.github.cyanheads/ensembl-mcp-server is a MCP server for the Idea phase that queries Ensembl for genes, sequences, variant consequences, orthologs, and cross-references through your agent.
What is this MCP server?
- Gene lookup and sequence retrieval against Ensembl REST (default https://rest.ensembl.org, overridable for GRCh37 or mir
- Variant consequence prediction and cross-reference (xref) resolution from the agent thread
- Ortholog discovery for comparing genes across species in planning conversations
- stdio via npm @cyanheads/ensembl-mcp-server (Bun runtimeHint) plus hosted streamable-http remote
- Configurable MCP_LOG_LEVEL and ENSEMBL_BASE_URL for local or alternate endpoints
- Server version 0.1.3
- npm package @cyanheads/ensembl-mcp-server
- Default Ensembl REST base https://rest.ensembl.org
What problem does it solve?
Agents hallucinate gene symbols and variant annotations when they cannot call Ensembl REST in real time.
Who is it for?
Indie builders prototyping genomics APIs, research copilots, or variant-explainer features who already reason in Ensembl terms.
Skip if: Clinical diagnostic workflows, bulk offline VEP at scale, or teams forbidden from sending queries to public genomics APIs.
What do I get? / Deliverables
After you register the server, your agent returns Ensembl-backed IDs, sequences, consequence predictions, and ortholog links you can paste into specs, schemas, and integration code.
- Agent-callable Ensembl gene, sequence, variant, ortholog, and xref results
- Configurable local stdio or remote HTTP MCP connection (v0.1.3)
- REST-aligned responses suitable for specs, tests, and integration stubs
Recommended MCP Servers
Journey fit
Canonical shelf is Idea because builders reach for authoritative genomic references before they commit to schemas, APIs, or product scope—not because Ensembl is only useful in discovery. Research fits first-pass gene IDs, cross-species orthologs, and variant consequence checks that inform what to build or validate.
How it compares
MCP bridge to Ensembl REST, not a local genome browser or a packaged bioinformatics skill library.
Common Questions / FAQ
Who is io.github.cyanheads/ensembl-mcp-server for?
Solo and small-team builders using Claude Code, Cursor, or similar agents who need authoritative Ensembl gene, sequence, variant, and ortholog data while designing or coding genomics features.
When should I use io.github.cyanheads/ensembl-mcp-server?
Use it during early research and integration design whenever the agent must resolve gene IDs, fetch sequences, predict variant consequences, or map orthologs instead of guessing from training data.
How do I add io.github.cyanheads/ensembl-mcp-server to my agent?
Add the npm package @cyanheads/ensembl-mcp-server to your MCP config with stdio transport (run start:stdio), set ENSEMBL_BASE_URL if you need GRCh37 or a mirror, optionally tune MCP_LOG_LEVEL, or point the client at the hosted streamable-http remote URL from server.json.