API reference

Document Parser API

Extract text and metadata from HTML, plain text, Markdown, CSV/TSV, JSON, XML and PDF. Read visible text from one static JPEG, PNG or WebP image up to 2 MB and 16 megapixels, including TikTok/Instagram post covers via data.cover_ocr. Empty PDF extraction is uncharged. Office files, archives and unsupported image formats return 422.

Read as Markdown Use with an AI agent

Endpoint

GET/v1/web/parselive · proven2 credits

Try in Playground

Parameters

Query parameters

Name	Required	Description	Example
`url`	yes	Public document URL. Also accepts one static JPEG, PNG or WebP image up to 2 MB and 16 megapixels for text extraction; other slides or video frames require separate requests.	`https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf`
`timeout`	no	Fetch budget in milliseconds, clamped to 1000-15000	`10000`
`dry_run`	no	Read a zero-credit estimate without fetching sources, running AI, or reserving credits. Cache status is a snapshot, not a guarantee at execution.	`1`

Examples

Make the request

curl

curl "https://www.monocrawl.com/v1/web/parse?url=https%3A%2F%2Fwww.w3.org%2FWAI%2FER%2Ftests%2Fxhtml%2Ftestfiles%2Fresources%2Fpdf%2Fdummy.pdf" \
  -H "x-api-key: mn_your_key_here"

TypeScript

typescript

const key = process.env.MONOCRAWL_API_KEY;
if (!key) throw new Error('Set MONOCRAWL_API_KEY on your server.');
const res = await fetch(
  "https://www.monocrawl.com/v1/web/parse?url=https%3A%2F%2Fwww.w3.org%2FWAI%2FER%2Ftests%2Fxhtml%2Ftestfiles%2Fresources%2Fpdf%2Fdummy.pdf",
  { headers: { "x-api-key": key } },
);
const body = await res.json();

if (!body.success) {
  // one error shape for every endpoint — see /docs/errors
  throw new Error(`${body.error.type}: ${body.error.message}`);
}

console.log(body.data, "credits left:", body.credits_remaining);

Python

python

import os
import requests

res = requests.get(
    "https://www.monocrawl.com/v1/web/parse?url=https%3A%2F%2Fwww.w3.org%2FWAI%2FER%2Ftests%2Fxhtml%2Ftestfiles%2Fresources%2Fpdf%2Fdummy.pdf",
    headers={"x-api-key": os.environ["MONOCRAWL_API_KEY"]},
    timeout=60,
)
body = res.json()

if not body["success"]:
    # one error shape for every endpoint — see /docs/errors
    raise RuntimeError(f"{body['error']['type']}: {body['error']['message']}")

print(body["data"], "credits left:", body["credits_remaining"])

Response

Response fields and example

This example is illustrative, not a captured live response. Variable-cost operations may settle a charge different from the list price below. A successful response puts the platform payload in data and reports the exact credits used, remaining balance, request id and cache status beside it.

200 · application/json

{
  "success": true,
  "platform": "web",
  "endpoint": "/v1/web/parse",
  "data": {
    "detected_type": "image",
    "text": "Example visible text",
    "texts": [
      "Example visible text"
    ],
    "truncated": false
  },
  "credits_used": 2,
  "credits_remaining": 99,
  "request_id": "req_…",
  "cached": false
}

The example is illustrative and shows a documented subset of data. The fields below are optional across supported sources; nullable fields can also be absent. Preserve unknown values and accept additional fields.

Schema basis: reviewed response shaper. See schema coverage and validation limits.

Download the data JSON Schema. Validate response.data, not the whole envelope. A valid shape does not establish that every field or source record was returned.

Field inside data	Type	Meaning
`url`	`string \| null`	url
`final_url`	`string \| null`	final url
`content_type`	`string \| null`	content type
`detected_type`	`string \| null`	detected type
`detected_via`	`string \| null`	detected via
`text`	`string \| null`	text
`byte_size`	`number \| null`	byte size; null is unknown.
`text_length`	`number \| null`	text length; null is unknown.
`pages`	`number \| null`	pages; null is unknown.
`metadata`	`object \| null`	Nested response fields; optional unless explicitly documented.
`truncated`	`boolean \| null`	Whether returned document text was cut short.
`texts`	`array`	Nested response fields; optional unless explicitly documented.
`_warnings`	`array`	Nested response fields; optional unless explicitly documented.

Static JPEG, PNG and WebP: one image up to 2 MB and 16 megapixels, normalized to fit 1024 pixels. Other carousel slides and video frames require separate requests. OCR can be inaccurate; output truncation fails rather than silently returning an incomplete transcription.

HTML, text and PDF parsing remain available without a model call. The existing endpoint price applies.

Failures use the typed error envelope. A confirmed uncharged or refunded failure reports zero; a pending reconciliation can report an unknown charge. Read credits_used and error.details.billing_status, and keep request_id for recovery. Response contract · Error reference

Document Parser API

Query parameters

Make the request

Response fields and example

More Web Scraping reference

First call in under a minute