Parse

Documents in, clean markdown out

Upload a PDF, Word file or spreadsheet and get structured markdown back. Text PDFs are parsed directly, scans go through OCR, and every conversion gets a quality score.

Scrape Crawl Map Search Parse Monitor Crawl tree

PDF + Office
Also EPUB, images and audio, all into the same markdown
OCR
Scanned pages with no text layer are read by OCR in the console and API
Audit
ROUGE-1, Jaccard, vocabulary coverage and length fidelity per file

A parser you can check

Text PDFs

Headings, lists and tables come through as markdown, page order kept.

Scans and photos

Pages with no text layer are detected and sent to OCR rather than returned empty.

Office files

Word, Excel, ODT and RTF convert to the same markdown as web pages.

A conversion audit

Each file is scored against its source, so a bad conversion is a number, not a surprise.

Searchable at once

Parsed files are chunked and embedded into a knowledge base with your web pages.

Not kept by the demo

The free tool converts in memory and stores nothing. Files you upload to the API stay in your knowledge base until you delete them.

How parsing works

01 / 03
Upload a file
POST /api/jobs/upload with the file and a knowledge base.
02 / 03
We detect and convert
Text layers are parsed directly; scanned pages go through OCR in the language you choose.
03 / 03
Check and search
Read the markdown and its audit score in the console, then search it with everything else.

Parse from your code

Uploads are multipart. The job returns the markdown, the chunks, and the audit when it completes.

curl -X POST https://crawl.sarv.com/api/jobs/upload \
  -H "x-api-key: YOUR_API_KEY" \
  -F "kb_id=kb_abc123" \
  -F "file=@./report.pdf"

Upload API reference

Questions

Can it read scanned PDFs?

Yes, in the console and the API: pages with no text layer are rendered and read with OCR. The free demo above handles text PDFs and tells you when a file needs OCR.

Which file types are supported?

PDF, Word (DOC, DOCX), Excel (XLS, XLSX), ODT, RTF, EPUB, images and audio through the API; the free demo takes PDF and office files up to 10 MB.

How do I know a conversion is good?

Each file gets an audit with ROUGE-1, Jaccard similarity, vocabulary coverage and length fidelity against its source, visible in the console.

Do you keep the files?

The free demo converts in memory and keeps nothing. Files uploaded through your account are stored in your knowledge base until you delete them, and we do not train on them.

Turn your documents into search

1 billion tokens free on sign-up, no card. Roughly 500 tokens a page.

Start with 1 billion free tokens Read the API docs