neosun100/deepwiki-scraper — reverse-engineered prompt

Build me a simple tool that can take a DeepWiki project link, or just a repo name like owner slash project, scrape all the documentation pages, and combine everything into one clean Markdown file that I can download.

I want both a nice web page and a command line option. The web page should feel polished, with a dark purple gradient look, language switching for English, Simplified Chinese, Traditional Chinese, and Japanese, a place to enter one or more Firecrawl API keys, and it should remember those keys in my browser so I don’t have to paste them every time. While it runs, show clear progress, page counts, and status messages.

Please add smart caching so if I scrape the same project again it can offer to reuse the saved result instead of calling the API again. Let me adjust concurrency and request delay, rotate multiple API keys, retry failures, and avoid rate limits. Make it easy to run locally, and also include Docker setup. Look up current docs online if you need to.