OLS4 Ontology Lookup Service MCP
An EMBL-EBI MCP endpoint for ontology class search, ancestors and descendants, including semantic embedding search.
- 108 ★ on the repository
- Last commit
- Apache-2.0
Minimal access
Why this level
- The tools only read public ontologies, but requests go to an external EMBL-EBI service.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/ols4-ontology-lookup-serviceDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The tools only read public ontologies, but requests go to an external EMBL-EBI service.
This is third-party code. Review the repository files before installing.
What it does
Ontology Lookup Service (OLS4) from EMBL-EBI keeps current versions of biomedical ontologies. Besides a website and REST API it has a public MCP endpoint, `https://www.ebi.ac.uk/ols4/api/mcp`, over Streamable HTTP. It runs without sessions: a client initializes and calls tools with POST requests, and listening GET requests return 405. Tools search classes by string, return ancestors, direct children and all descendants of a class, list ontologies, and search by meaning using vector embeddings. Before semantic search you fetch the list of embedding models. There are also OpenAI-compatible `search` and `fetch` tools with an ID of the form `go+IRI`.
Who it is for. Bioinformaticians and researchers who need an agent to map terms to ontologies such as GO or EFO.
Good fit when
- Free text in a table is mapped to ontology terms.
- All descendants of a class are needed to widen a database query.
- A disease name is searched by meaning when no exact phrase exists in the ontology.
- An agent lists the available ontologies before choosing a vocabulary.
Not a fit when
- You need your own OLS instance with private data: the public endpoint serves shared ontologies.
- A long session with server notifications is required: it does not support them.
- You work offline.
Example request
Find GO classes about cellular respiration and list all descendants of the most general one.MCP
- Transport
- http
- Authentication
- not required
Limitations
The public endpoint runs on EMBL-EBI infrastructure and the README sets no terms for heavy load: requests go to an external service. The sessionless endpoint sends no server messages, so clients that rely on event subscriptions will not fit. Semantic search needs a model with the `can_embed` flag, picked from the list. The repository holds three projects (a data loader, an API server and a React frontend), and deploying your own instance needs Docker or Kubernetes.
How to disable. Remove the server entry from the MCP client settings. Stop your own OLS instance with Docker Compose or Kubernetes commands.
FAQ
What is the MCP address?
`https://www.ebi.ac.uk/ols4/api/mcp`, Streamable HTTP transport without sessions.
Why does GET return 405?
The server keeps no SSE connections so abandoned ones cannot exhaust capacity. Working requests use POST.
How do I search by meaning?
Call the embedding model list first, pick a model with `can_embed=true` and pass a natural-language query.
What is the ID format for fetch?
A string `ontologyid+entityIri`, for example `go+http://purl.obolibrary.org/obo/GO_0008150`.
How do I deploy my own OLS?
Run the data load with `dataload.sh` and an ontology config, then `docker compose up`. Details are in the repository README.
Discussion
No questions or reviews yet.
Related
MCP serversAn MCP server from the SIB institute gives an agent query examples and class schemas for SPARQL to UniProt and Bgee.
A medical code reference for your agent: search, map and navigate SNOMED CT, ICD-10, RxNorm and LOINC concepts across vocabularies
The official Open Targets MCP server: an agent searches genes, diseases and drugs and queries target-disease associations
Open Ontologies
open-ontologies
Planning changes to OWL ontologies: impact assessment, rollback and verifiable evidence for a reviewer. A single Rust binary for a narrow audience
NVIDIA Auto Ontology semantic layer as an MCP server: SQL gets validated, facts land in a personal knowledge base.
An MCP server for searching and downloading academic papers from arXiv, PubMed, bioRxiv, Semantic Scholar and other open sources
README in short
The README points to the website, REST API and MCP endpoint, then covers the three repository projects and two deployment paths: Docker Compose and Kubernetes.