Pagination runner

Use this when you have API docs but do not know which fields to enter. Generate a prompt, send it to Claude, ChatGPT, or another AI, then paste the returned JSON object back here. The app fills the values the AI provided and clears untouched defaults it did not provide.

Enter the upstream API endpoint exactly as the API expects it. If a credential must live in the URL, use a placeholder such as {{key}}; Clay will pass the value on the generated runner URL at run time.

Choose the pagination style from the API docs. Max pages is a safety cap, not the number of pages to fetch; the runner stops automatically when the API has no next page.

Return the full item by default, flatten JSON:API attributes for easier Clay columns, or select only the fields Clay needs.

These controls prevent runaway loops and help with rate-limited APIs. The defaults stop on empty pages and repeated next links while still allowing full pagination.

List only the header names Clay should forward at run time. Use test credentials here to verify the runner; test credentials and query values are not saved.

{}

Review the tested configuration, then save. Saved runners are locked so live Clay tables keep a stable URL and behavior.