How to Configure MCP Tool Parameters in a Scenario

If you want the AI assistant to pass data when calling a section (e.g. city, date, identifier), you need to define these parameters in the MCP Tool settings and use them in the section logic. This guide describes only the actions in the scenario and in the Constructor — no implementation details.

When You Need This

  • You need the AI to pass values when calling the tool (e.g. city for weather, currency for exchange rate).
  • The section is already set up as an MCP Tool and you want to add or change input parameters.
  • You need to use the passed values in messages, conditions, or actions within the section.

What You Need to Know

  • Input Schema — the list of tool parameters (name, type, required, description). You fill it in the MCP Tool trigger settings on the section’s Start node.
  • Parameters you add to the Input Schema are filled with values from the AI client when the MCP tool is called.
  • These values are available in the scenario: you can use them in message texts, conditions, and action parameters the same way as other scenario variables (via placeholders by parameter name).

Before You Start

Step-by-Step Instructions

1. Add or Change Parameters in Input Schema

  1. Go to the Scenario module via Menu -> Scenario, open the bot’s scenario, and select the section that is the MCP tool.
  2. Click the section’s Start node (entry point) — the settings panel opens on the right.
  3. In the MCP Tool trigger settings block, find the Input Schema section (input parameters schema).
  4. Add a parameter: use the add-parameter button (if available) or fill in the form for a new parameter.
  5. For each parameter specify:
    • Name — parameter name in Latin (e.g. city, currency, date). You will use this name to reference the value in the scenario.
    • Type — data type (string, number, boolean, etc.) so the AI client and platform validate values correctly.
    • Required — whether this parameter must be passed when the tool is called.
    • Default (optional) — value used if the AI does not pass the parameter.
    • Description (optional) — short description for the AI so it knows what to pass.
  6. Repeat steps 4–5 for all parameters the tool should accept.
  7. Save the scenario changes.

After saving, the platform exposes these parameters to the AI client; when the tool is called, the passed values are available in the scenario.

2. Use Parameters in the Scenario

Values passed by the AI when calling the tool are available within that section. You can use them anywhere the scenario editor supports variable placeholders.

  1. In message texts — in Message nodes you can add a placeholder for the parameter value by its name from the Input Schema (placeholder format depends on the editor; usually a field with hints or a list of available variables).
  2. In conditions — in condition nodes you can compare the parameter value with a constant or another variable and control branching.
  3. In actions — in action parameters you can substitute the parameter value (e.g. pass it in a request URL or in the action body).

Tip: parameter names in placeholders must match those in the Input Schema. If the parameter is named city, reference it by that same name in the scenario.

3. Verify It Works

  1. In the AI client (e.g. Cursor) call the tool and pass test values for the parameters.
  2. Check that the section response contains the expected data and that the passed values appear in messages or results (if you used them there).
  3. If values do not appear — verify that parameter names in the Input Schema and in scenario placeholders match and that scenario changes were saved.

What Next

  • Add or update parameters in the Input Schema when you change the tool logic so the AI always passes the right data.
  • Use parameter descriptions in the Input Schema — this helps the AI client form correct calls.