For the complete documentation index, see llms.txt. This page is also available as Markdown.

Custom

Connect any MCP-compatible LLM tool to Data Product MCP.

Connect any MCP-compatible LLM tool to Data Product MCP using streamable HTTP.

What you need

Your DataOS instance URL, a client that supports streamable HTTP, and — if your client uses the API token method — a DataOS API token. See API tokens for how tokens work.

Connect

Configure your client with these values:

Field
Value

Server name

DataProduct-MCP

Transport

Streamable HTTP

URL

https://<instance-url>/mcp/api/v1

Leave headers empty. Your client should prompt you to authorize in a browser window when you connect.

If your client asks for OAuth settings instead of discovering them, use these values:

Setting
Value

Authorization URL

https://<instance-url>/mcp/api/v1/oauth/authorize

Token URL

https://<instance-url>/mcp/api/v1/oauth/token

Client registration URL

https://<instance-url>/mcp/api/v1/oauth/register

Supported scopes

openid profile email groups federated:id offline_access

PKCE

Required, with the S256 method. PKCE is a sign-in security check that most OAuth clients and libraries handle automatically. If your client asks for a code challenge method, select S256.

Client secret

Not required

Clients that support dynamic client registration get a client ID automatically from the client registration URL.

Field
Value

Server name

DataProduct-MCP

Transport

Streamable HTTP

URL

https://<instance-url>/dataproduct-mcp/api/v1

Header key

apikey

Header value

<API_TOKEN>

You can copy these from Home → Connect to Data Product MCP in DataOS: open the config modal and select Copy. Then open your tool's MCP settings, add a server with the values above, save, and connect.

Verify

Ask What data products am I authorized to access?. A working connection returns products scoped to your token and permissions. Next: Discover in natural language.

Last updated

Was this helpful?