Claude

Anthropic’s Claude connects to the Datamesh MCP server in three places: Claude.ai on the web, the Claude Desktop app, and the Claude Code CLI. The web and desktop apps use the hosted server over OAuth; Claude Code can use either the hosted server (via a header) or the local stdio server.

Claude.ai (web)

Claude.ai reaches the hosted server through a custom connector and signs in with your Oceanum.io account.

1. In Claude’s settings, open Connectors and choose Add custom connector.

2. Give it a name (for example Oceanum Datamesh) and set the URL to https://mcp.oceanum.io/datamesh.

3. Open Advanced settings and set the OAuth Client ID to:

   g3GYWWPrj1LKB89bhsYrZ8kebg5arKvu

   The value is a public PKCE client ID with no secret — it is safe to enter.

4. Save the connector, then click Connect and log in with your Oceanum.io account.

Once connected, Claude runs read-only as your user. query_data results come back inline (capped at ~50 MB), so large results can’t be returned in the chat — but Claude can call export_query, which on the hosted connector returns a time-limited download link to the full result for you to fetch (the link needs no token, so treat it as a secret). To have export_query write results straight to a file instead, use the local server below.

Claude Desktop

Claude Desktop supports the same OAuth connector as the web app, and also a local stdio server for full read/write.

Custom connector (OAuth)

Follow the same steps as Claude.ai above: open Settings → Connectors → Add custom connector, set the URL to https://mcp.oceanum.io/datamesh, set the OAuth Client ID under Advanced settings to g3GYWWPrj1LKB89bhsYrZ8kebg5arKvu, then Connect and sign in.

Local server (stdio)

For full read/write, and to have export_query write results straight to a file on your machine (rather than return a download link), run the server locally. In Claude Desktop, open Settings → Developer → Edit Config to open claude_desktop_config.json, and add:

{
  "mcpServers": {
    "oceanum-datamesh": {
      "command": "uvx",
      "args": ["oceanum-mcp", "datamesh"],
      "env": { "DATAMESH_TOKEN": "your-token-here" }
    }
  }
}

Replace your-token-here with your Datamesh token and restart Claude Desktop. This requires uv on your PATH.

Claude Code (CLI)

The Claude Code CLI adds MCP servers with claude mcp add. Use whichever transport suits you.

Hosted server over a header token:

claude mcp add --transport http oceanum-datamesh https://mcp.oceanum.io/datamesh --header "X-DATAMESH-TOKEN: <datamesh-token>"

Local stdio server (full read/write):

claude mcp add --transport stdio oceanum-datamesh -- uvx oceanum-mcp datamesh

The local server authenticates with the DATAMESH_TOKEN environment variable, so export it in the shell where Claude Code runs:

export DATAMESH_TOKEN="your-token-here"

Keep your token safe

Your Datamesh token grants access to all of your organisation’s permissioned datasources — including the ability to modify or delete them. Treat it as a secret, and prefer the hosted read-only connector for exploratory use.