URL and PDF to Markdown
Pricing
Pay per usage
URL and PDF to Markdown
Turn public web pages and PDF links into clean Markdown plus title, description, and word count. Built for LLM and RAG ingestion.
Pricing
Pay per usage
Rating
0.0
(0)
Developer
Ross AI
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
3 days ago
Last modified
Categories
Share
This actor fetches public web pages and direct PDF links and returns clean Markdown for LLM or RAG ingestion. HTML goes through a reader view, then to Markdown. PDFs are read as text. Each row also has the final URL, source type, title, description, and word count.
A page that fails, or that has no extractable text, is one error row. The rest of the batch still finishes.
Input
{"urls": ["https://example.com","https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf"]}
| Field | Default | What it does |
|---|---|---|
urls | required | 1 to 30 public http(s) pages or PDF links. Duplicates are dropped. |
Responses larger than 8 MB are skipped. Markdown is cut at 150,000 characters, and truncated is set when that happens.
Output
{"url": "https://example.com","finalUrl": "https://example.com/","sourceType": "html","title": "Example Domain","description": null,"markdown": "Example Domain\n\nThis domain is for use in illustrative examples.","wordCount": 12,"truncated": false,"status": "ok","error": null}
sourceType is html or pdf. A file is treated as a PDF when the content type says so, the path ends in .pdf, or the body starts with %PDF-. The SUMMARY record counts HTML and PDF successes.
Proxy and limits
The actor asks for the included datacenter proxy group BUYPROXIES94952. It does not use residential, Google SERP, or unblocker proxies. If that group is unavailable, it continues without a proxy. A page that answers 403 or 429, or times out, through the datacenter address is retried once with no proxy.
Localhost, link-local, and private addresses are refused. Each redirect is checked again before it is followed.
Pages that only render in a browser, and PDFs that are scans without a text layer, come back empty. This actor does not run a headless browser.
Pricing
Pay-per-event code is in the actor and is not enabled. The event name is page-converted at $0.001 per page that produced Markdown, with no start fee. Error rows are not billed. Until monetization is turned on, runs are not charged for results. See PRICING.md.