Skip to main content

Connect Other Clients

Any client that speaks the Model Context Protocol can use UbiMCP. This page collects setup snippets for the most common ones. If your client is not listed, the pattern is the same: point it at https://mcp.unibiointelligence.com/mcp over HTTP.

Cursor​

Cursor reads MCP servers from ~/.cursor/mcp.json:

{
  "mcpServers": {
    "ubi-mcp": {
      "url": "https://mcp.unibiointelligence.com/mcp"
    }
  }
}

Reload Cursor after editing.

VS Code with GitHub Copilot Chat​

Recent versions of GitHub Copilot Chat support MCP through workspace or user settings. Add to settings.json:

{
  "chat.mcp.servers": {
    "ubi-mcp": {
      "url": "https://mcp.unibiointelligence.com/mcp"
    }
  }
}

Continue.dev​

In ~/.continue/config.json:

{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "http",
          "url": "https://mcp.unibiointelligence.com/mcp"
        }
      }
    ]
  }
}

Anthropic SDK (Python)​

from anthropic import Anthropic

client = Anthropic()

response = client.messages.create(
    model="claude-opus-4-5",
    max_tokens=4096,
    mcp_servers=[{
        "type": "url",
        "url": "https://mcp.unibiointelligence.com/mcp",
        "name": "ubi-mcp",
    }],
    messages=[{
        "role": "user",
        "content": "Use UbiMCP to look up TYK2 expression across tissues."
    }],
)
print(response.content)

Custom client over raw MCP​

UbiMCP is a standard MCP server. Any library that implements the protocol (the official mcp Python SDK, the TypeScript SDK, or community implementations in Go, Rust, and others) can connect. Point the client at the HTTP endpoint and call initialize, then tools/list, then tools/call as usual.

Reference implementations: https://github.com/modelcontextprotocol.