Skip to main content

Overview

The LLM Search block sends a single prompt to several answer-engine platforms in parallel and returns each platform’s answer plus its citations. It also returns one combined, deduplicated list of citation URLs. Use it to see how AI engines answer a query and to gather source URLs for further research.

Configuration

Prompt (Required)

The question or instruction to send to each platform. Example:
You can reference placeholders from previous steps: {{step_1.output}}

Platforms

Select which answer engines to query. They run in parallel. Select at least one. Default: All platforms

Country

A 2-letter country code for localized results. Default: United States (us) Options: United States, United Kingdom, Canada, Australia, India, Germany, France

Error Handling

Define what happens if the block fails. Options:
  1. Terminate Workflow: Stop execution immediately.
  2. Continue Execution: Proceed to the next step despite the error.
If a single platform fails, it returns an empty answer with an error field while other platforms still return results.

Output Data

Returns a JSON object.

prompt

The rendered prompt that was sent.

platforms

The list of platforms that were queried.

results

An object keyed by platform. Each entry has an answer and a citations array of { url, title }. A failed platform also includes an error field.

citation_urls

A combined, deduplicated list of citation URLs from all platforms, ready for scraping. Accessing fields: {{step_1.output.results.chatgpt.answer}}, {{step_1.output.citation_urls[0]}}

Sample Output


Best Practices

  • Write a specific prompt; vague prompts return generic answers across engines.
  • Select only the platforms you need to keep runs fast and focused.
  • Use citation_urls to feed a Loop and Web Scrape for deeper research.
  • Set the Country to match the market you are studying.
  • Compare answers across platforms to spot consensus and gaps.

Common Use Cases


What’s Next