Connecting¶
The InterPro MCP server is hosted by EMBL-EBI and speaks MCP over Streamable HTTP. You don’t need to install, run or configure anything: just point your AI client at the public endpoint:
Endpoint
https://www.ebi.ac.uk/interpro/mcp
It’s a public, read-only service: no account, API key or authentication is needed. Once your client is connected, ask questions in plain language; the assistant discovers the tools and picks them itself.
The steps below cover the common clients.
Claude Code (CLI)¶
Add the server with one command:
claude mcp add --transport http interpro https://www.ebi.ac.uk/interpro/mcp
By default this registers it for the current project only; add -s user to make it available in
all your projects:
claude mcp add -s user --transport http interpro https://www.ebi.ac.uk/interpro/mcp
Check the connection, then use it:
claude mcp list # interpro -> ✓ Connected
Inside a session, /mcp lists the server and its tools; the model sees them namespaced as
mcp__interpro__getEntry, mcp__interpro__searchInterPro, ….
Claude Desktop & claude.ai¶
Open Settings → Connectors, choose Add custom connector,
give it a name (e.g. InterPro) and paste the URL:
https://www.ebi.ac.uk/interpro/mcp
Then enable it from the tools / connectors menu in a chat, and the InterPro tools become available to the model.
Other MCP clients¶
Any client that supports the Streamable HTTP transport can point straight at the URL. A typical configuration entry:
{
"mcpServers": {
"interpro": {
"type": "streamable-http",
"url": "https://www.ebi.ac.uk/interpro/mcp"
}
}
}
Check it’s working¶
Ask your assistant something that needs InterPro, for example:
“Using InterPro, what is the kringle domain and which entry describes it?”
If the server is connected the assistant will call a tool (e.g. searchInterPro) and answer
with an accession such as IPR000001. In Claude Code, /mcp also shows the live connection
status and the tool list.