jliyon23/crawlflow ? reverse-engineered prompt

Build me a visual web crawler app where I can draw a workflow and then run it to scrape data.

I want the editor to let me connect simple blocks like Input, Request, Parser, Follow links, Pagination, Loop, Branch, and Output. The flow should pass variables between blocks, so I can start from a URL, a CSV, or a small list of values, then make requests, extract fields from HTML or JSON, follow links, handle pagination, and save the final results as JSON, JSONL, or CSV.

It should also have a working backend that runs the workflow as a real crawler, with things like retries, headers, cookies, proxies, delays, robots.txt, and concurrency handled for me. Please include a validate command, a run command, and a local server so I can open the editor in the browser and test the sample workflows.

Use the example workflows as a guide, and if you need current docs for any libraries, look them up online.