Meta MCP
AI-generatedSummary
Calls an HTTP webhook URL with a chosen method (GET, POST, PUT, DELETE) for each input item and returns the JSON response, so a workflow can be exposed as a structured tool for AI Agents.
Inputs
- method (required) — HTTP method: GET (default), POST, PUT or DELETE.
- url (required) — Webhook URL to call. Leading/trailing whitespace and a leading '/' are stripped. Supports <
> placeholders, replaced with values from the injected Bonnie context (AsyncLocalStorage); missing/null values become empty strings. - authentication — None (default), Bearer Auth (httpBearerAuth credential, sent as 'Authorization: Bearer
') or Header Auth (httpHeaderAuth credential, sent as a custom header name/value). - bodyJson (required) — JSON body, shown only for POST and PUT. Supports <
> placeholders and is parsed with JSON.parse, so invalid JSON causes an error. Defaults to {}. - meta — Additional structured data/schema describing the tool. Declared in the UI but not read by the execute code.
- curlImport — UI helper for importing a cURL command to fill in the fields.
Output shape
A list of items, with one item per response object. If the response is a JSON array, each element becomes its own item. Otherwise the whole response becomes a single item's json.
Sends an Accept: application/json header and requests a JSON response. A body is sent only for POST/PUT. With Continue On Fail enabled, a failed item yields {json:{error:message}}. Otherwise the node throws a NodeApiError.
Examples
Example 1: Trigger an external n8n webhook workflow from an AI Agent with a JSON payload
Method=POST, URL=https://example.com/webhook/abc, Authentication=Header Auth, Body={"query":"<
Example 2: Fetch data from a protected endpoint
Method=GET, URL=https://api.example.com/items, Authentication=Bearer Auth