Connection¶

The Connection tab is where you configure how the API Connector calls your LLM API.

API Endpoint¶

Set the HTTP method and URL for your API:

• Method — POST (default) or GET
• URL — the full API endpoint URL (e.g., https://api.example.com/v1/chat/completions)

The URL supports {{variable}} placeholders that will be replaced with dataset values.

Headers¶

Configure HTTP headers for authentication and content type:

Click + Add header to add additional headers. Headers also support {{variable}} placeholders.

Query Parameters¶

Add URL query parameters as ?key=value pairs. These are appended to the API endpoint URL and also support {{variables}}.

Request Body Template¶

Write the JSON body that will be sent with each API request. Use {{column_name}} placeholders to inject values from your dataset.

{
  "messages": [
    {"role": "user", "content": "{{input}}"}
  ],
  "model": "gpt-4o-mini"
}

{
  "query": "{{input}}",
  "context": "{{context}}",
  "max_tokens": 1024
}

The placeholders match column names from your dataset (configured in the Columns tab).

Test Connection¶

Click Send Test Request to verify your API configuration works. The response will be available in the Response tab for mapping.