Connections

Connect an OData service

A connection is the foundation of everything else. You register a service once, DataTether reads its metadata, and you get typed MCP tools you can review before anyone uses them. This page walks through each step and shows exactly what you get back.

  1. 1

    Open Add Connection

    Console -> Add Connection

    From the console, choose Add Connection to start a new connection. This is the same flow you will use later for SAP Gateway, SAP S/4HANA, or any other OData v2 or v4 endpoint.

    What you'll see

    The Add Connection form, ready for a new connection.

    This is where the product screen for this step will appear.

  2. 2

    Enter the service URL and choose auth

    Add Connection -> Connection

    Provide the service document URL and tell DataTether how to authenticate. Three auth types are supported:

    • No Auth for public or demo services such as the Northwind sample.
    • Basic for a service user with a username and password held in the credential vault.
    • OAuth for token-based access, with a token endpoint and scope.
    What you'll seeConnection saved
    Service URL : https://services.odata.org/V2/Northwind/Northwind.svc/
Auth type   : No Auth
Status      : reachable
  3. 3

    Discover the metadata

    Add Connection -> Discovery

    DataTether reads the service metadata document and turns it into a capability inventory you can configure. Nothing is exposed yet. This step is read-only and safe to run against production.

    What you'll seeMetadata parsed

    A structured inventory of the service:

    • Entity sets and entity types, with their keys and properties
    • Navigation properties between related entities
    • Function and action imports the service exposes
  4. 4

    Generate tools

    Add Connection -> Discovery

    For each entity you keep, DataTether generates a typed set of MCP tools. Read tools are created first; write tools are optional and stay off until you turn them on.

    What you'll seeTools generated
    list_SalesOrder      filter and page through records
get_SalesOrder       fetch one record by key
count_SalesOrder     count records with filters
search_SalesOrder    free-text search across fields
create_SalesOrder    create a record   (optional, off by default)
update_SalesOrder    update a record   (optional, off by default)
delete_SalesOrder    delete a record   (optional, off by default)
  5. 5

    Configure exposure

    Add Connection -> Configure

    Decide what an AI client is actually allowed to see and do. For each connection you can:

    • Select which entities and fields are exposed, and hide the rest
    • Choose which operations are allowed, or set the whole connection to read-only
    • Rename technical operations into business-readable tool names
    • Require human confirmation before any write runs
    What you'll see

    Exposure settings: selected entities, fields, and confirmation rules.

    This is where the product screen for this step will appear.

  6. 6

    Publish the connection

    Add Connection -> Publish

    Publishing makes the connection available as a hosted MCP endpoint. Your selected tools, resources, and prompts are now discoverable by any MCP-compatible client over Streamable HTTP, with SSE compatibility for older clients.

    What you'll seeLive endpoint
    POST https://mcp.suriz.in/{org}/{system-name}/mcp

tools/list       list the tools you exposed
tools/call       run a tool with arguments
resources/list   list schemas and documentation
resources/read   read a single resource

Related