Configuration

There are two independent aspects of configuration for your InfoTrack MCP integration: how your agent discovers services at runtime, and which services are available to your connection.

Runtime Discovery Mode

When your agent connects to the Gateway, it needs to know what services are available. How this discovery happens depends on your agent's capabilities and how many services you have enabled.

Standard Mode

The full list of available services is presented to your agent on connection. The agent sees every service and can invoke any of them directly. This is the default and works well with capable models that can reason across a broad set of options.

Search and Invoke

If your agent has limited context capacity, or you have a large number of services enabled, the Gateway can expose a lightweight search-first pattern instead. Only a single discovery tool is presented initially. The agent describes what it needs, the Gateway identifies the right service, and the agent then invokes it.

This frees up context for the model and is recommended when working with smaller or more cost-sensitive models, or when the number of available services would otherwise consume significant token budget.

Which mode should I use?

Start with Standard Mode. If you notice your agent struggling with service selection or you want to reduce token usage, ask your InfoTrack representative to switch to Search and Invoke.

Available Services

Separately from how your agent discovers services, you can configure which services and capabilities are available to your integration. This is managed by your InfoTrack representative based on your organisation's needs and the products available to your account.

We recommend starting with the full catalogue so your agent has access to all available services. If you need to restrict access to a specific subset, your representative can configure this for you.

For a full list of services currently available through the Gateway, see the Available Capabilities page.