Document extraction  /  API {}

DocParse

Upload a file or pass a URL. Get text, per-page text, metadata and page counts back as JSON. Built on unpdf inside a Worker: no queue, no cold start bill, no 30 second timeout.

checking 7 endpoints Cloudflare edge
docparse · live
$ curl -s -X POST "https://docparse.ayamk.in/v1/pdf/text" -H "content-type: application/json" -d '{"url":"https://arxiv.org/pdf/1706.03762v7"}' | jq '.pageCount, .words'
{
  "pageCount": 15,
  "characters": 41288,
  "words": 6842,
  "truncated": false,
  "metadata": { "title": "Attention Is All You Need", "producer": "pdfTeX" },
  "pages": [
    { "page": 1, "text": "Provided proper attribution is provided..." },
    { "page": 2, "text": "1 Introduction Recurrent neural networks..." }
  ...
01 - ENDPOINTS

7 endpoints, one key

Every route below is described in the OpenAPI document. Import it into Postman or generate a client.

MethodPathWhat it does
POST/v1/pdf/textFull text plus per-page text, page count and metadata.
POST/v1/pdf/infoPage count, metadata, encryption flag. Cheapest call.
POST/v1/pdf/pagesExtract only the pages you name: 1,3,5-7.
POST/v1/docx/textExtract text, headings, lists and tables from a .docx.
POST/v1/html/textReadable article text from a URL or raw HTML.
POST/v1/html/markdownHTML to clean markdown.
GET/v1/formatsSupported inputs and limits.
02 - RESPONSE

What you actually get back

curl
curl -s -X POST "https://docparse.ayamk.in/v1/pdf/text" -H "content-type: application/json" -d '{"url":"https://arxiv.org/pdf/1706.03762v7"}' | jq '.pageCount, .words'
200 application/json
{
  "pageCount": 15,
  "characters": 41288,
  "words": 6842,
  "truncated": false,
  "metadata": { "title": "Attention Is All You Need", "producer": "pdfTeX" },
  "pages": [
    { "page": 1, "text": "Provided proper attribution is provided..." },
    { "page": 2, "text": "1 Introduction Recurrent neural networks..." }
  ],
  "durationMs": 626
}
Limit worth knowing. Text-based PDFs only. Scanned documents return little or no text because there is no built-in OCR.
03 - WHY

Built for the boring part

No job queue

Extraction happens in the request. A 15 page paper comes back in ~600ms, so your ingestion loop stays synchronous.

Two ways to send a file

multipart/form-data for direct uploads, or a JSON URL for files already in storage. Same response shape either way.

Bounded by design

8MB input, 100 pages per request, explicit truncation flags. Predictable cost, no surprise bills when someone uploads a 900 page scan.

04 - PRICING

Priced per request, not per seat

Basic
$0
25 requests / month
Card required by the marketplace. Enough to ship an integration.
Ultra
$75/mo
100,000 requests
Batch jobs and backfills.
Mega
$150/mo
400,000 requests
For platform-scale volume. Overage billed per request.

All plans are managed by the RapidAPI marketplace, which applies a marketplace fee before payout. Annual and custom limits are available on request.

05 - USE CASES

Where this earns its place

01

RAG ingestion

02

Invoice and contract parsing

03

Resume screening

04

Support ticket attachment reading

06 - FAQ

Questions we get asked

Can it read scanned PDFs?

It extracts whatever text layer exists. A pure image scan has none, so you get an empty string and a warning. Add your own OCR step if you need one.

What is the largest file I can send?

8MB or 100 pages per request, whichever comes first. Larger documents should be split or fetched page ranges with /v1/pdf/pages.

Do you keep my documents?

No. The body is processed in memory and discarded when the response is returned. Nothing is written to storage.

Does it handle password protected PDFs?

It reports them as encrypted with a 422 and a clear message. Decryption is not supported.

Subscribe to DocParse

Free tier included. The listing is published on the RapidAPI marketplace; the key works against every endpoint on this page.