PDF to Markdown Converter

Convert a PDF from a public URL or a file upload into clean Markdown. Files up to 256 MB are supported. The .md result is written to the run's default key-value store, together with a machine-readable OUTPUT record.

The Actor uses a fast local PDF parser that detects text-based and scanned documents, preserves reading order, and converts headings, lists, tables, links, and text formatting to Markdown. It can be called through the Apify API or from Apify Console.

Input

Field	Type	Default	Description
url	string	sample PDF	Public HTTP(S) URL of the PDF.
fileUpload	string	-	Uploaded PDF or Apify key-value store record URL. Takes precedence over url.

Only PDF input is accepted. The URL does not need to end in .pdf; the downloaded content is checked for a PDF header before conversion starts. Password-protected PDFs are rejected because their content cannot be read without the password.

Text must be embedded in the PDF. Image-only scans are not OCR'd; they produce an OUTPUT record with error.code = "NO_TEXT" and a suggestion to use an OCR-enabled pipeline.

Output

For an input named input.pdf, the Markdown document is stored in the run's default key-value store under input.md with content type text/markdown; charset=utf-8.

When the run finishes, the Output tab in Apify Console lists the Markdown document and links to the OUTPUT record. These links are defined in .actor/output_schema.json.

An OUTPUT JSON record describes the result:

{
  "success": true,
  "message": "Converted a 5.1 MB PDF (12 pages) to Markdown (2.3 MB). The result is stored as \"pptdemo1.md\" in the run's key-value store.",
  "input": {
    "source": "https://example.com/pptdemo1.pdf",
    "filename": "pptdemo1.pdf",
    "bytes": 5321132,
    "mb": "5.1 MB",
    "pages": 12
  },
  "output": {
    "key": "pptdemo1.md",
    "storeId": "STORE_ID",
    "filename": "pptdemo1.md",
    "contentType": "text/markdown; charset=utf-8",
    "bytes": 2411724,
    "mb": "2.3 MB",
    "downloadUrl": "https://api.apify.com/v2/key-value-stores/STORE_ID/records/pptdemo1.md"
  },
  "pricing": {
    "eventName": "pdf-mb",
    "unitPriceUsd": 0.001,
    "chargedEvents": 6,
    "estimatedCostUsd": 0.006,
    "charged": true,
    "actorStart": {
      "eventName": "apify-actor-start",
      "unitPriceUsd": 0.002,
      "events": 1,
      "estimatedCostUsd": 0.002,
      "chargedAutomatically": true
    },
    "estimatedTotalCostUsd": 0.008
  },
  "limits": {
    "memoryMb": 512,
    "maxMemoryMb": 512,
    "diskMb": 1024,
    "maxInputBytes": 268435456,
    "maxInputMb": 256
  }
}

On failure, success is false, error.code identifies the problem, and error.message contains a user-facing explanation. The OUTPUT record is written for both successful and failed runs.

Filename rules

• input.pdf -> input.md
• report.PDF -> report.md
• report -> report.md
• report.html -> report.md
• report.docx -> report.md
• report.tar.gz -> report.tar.md

Filenames are sanitized before use with a cross-platform filename cleaner. Directory components are removed, characters that are invalid in filenames are replaced, reserved device names are defused, and trailing dots or spaces are removed. The remaining name is normalized to letters, digits, ., _, and -, and truncated to a safe length. For example, hello world.pdf becomes hello_world.md.

Pricing

The Actor uses pay-per-event (PPE) pricing with two charges.

Per-run fee

Apify charges the synthetic apify-actor-start event automatically at $0.002 per start ($2 per 1,000 runs). It applies to every run, including runs that later fail, and is not charged by the Actor code.

The event is charged once for every run. Because the Actor's memory is fixed at 512 MB, each run incurs exactly one start event ($0.002).

Per-MB fee

The Actor charges the event pdf-mb once per started megabyte of the input file, at $0.001 per event:

• 1 byte to 1 MB -> 1 event -> $0.001
• 5.1 MB -> 6 events -> $0.006
• 128 MB -> 128 events -> $0.128
• 256 MB -> 256 events -> $0.256

This event is charged only after the Markdown document has been saved to the key-value store. Failed runs, including oversized, non-PDF, password-protected, and no-text inputs, are not charged for pdf-mb.

Total cost

Input	Start fee	pdf-mb	Total
1 byte to 1 MB	$0.002	$0.001	$0.003
5.1 MB	$0.002	$0.006	$0.008
128 MB	$0.002	$0.128	$0.130
256 MB	$0.002	$0.256	$0.258

The pricing object in OUTPUT reports both the pdf-mb charge and the automatic start fee.

Size limits

The Actor runs with 512 MB of memory and 1 GB of disk (Apify gives each run disk space equal to twice its memory). To leave headroom for PDF processing and the output file, the input limit is half the memory:

max input size = memory / 2 = 256 MB

Memory	Disk	Maximum input
512 MB (fixed)	1 GB	256 MB

The Actor checks the remote Content-Length before downloading. If the server does not report a size, the same limit is enforced while streaming the download. An oversized file produces an OUTPUT record with error.code = "FILE_TOO_LARGE" and no pdf-mb charge.

Files larger than 256 MB cannot be processed by this Actor; split them into smaller PDFs before uploading. The default run timeout is 10 minutes.

Calling the API

Start an asynchronous run:

curl -X POST "https://api.apify.com/v2/acts/OWNER~convert-pdf-to-markdown-api/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.pdfpdf.com/samples/pptdemo1.pdf"
  }'

The response contains data.id, data.defaultKeyValueStoreId, and data.status. Poll the run, then read the metadata and download the result:

curl "https://api.apify.com/v2/actor-runs/RUN_ID?token=$APIFY_TOKEN"

curl "https://api.apify.com/v2/key-value-stores/STORE_ID/records/OUTPUT?token=$APIFY_TOKEN"

curl -OJ "https://api.apify.com/v2/key-value-stores/STORE_ID/records/pptdemo1.md?token=$APIFY_TOKEN"

Synchronous call (waits for the run to finish):

curl -X POST "https://api.apify.com/v2/acts/OWNER~convert-pdf-to-markdown-api/run-sync?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com/input.pdf"}'

Uploading a file is also supported. Upload the PDF to a key-value store, then pass the record URL as fileUpload:

curl -X POST "https://api.apify.com/v2/acts/OWNER~convert-pdf-to-markdown-api/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "fileUpload": "https://api.apify.com/v2/key-value-stores/STORE_ID/records/input.pdf"
  }'

Error codes

Code	Meaning
INVALID_INPUT	Missing/invalid URL or unsupported scheme.
DOWNLOAD_FAILED	The source could not be downloaded (DNS, TLS, HTTP 4xx/5xx, timeout).
EMPTY_FILE	The source returned an empty file.
NOT_A_PDF	The file is not a valid PDF or has no readable pages.
PASSWORD_PROTECTED	The PDF is encrypted and needs a password.
NO_TEXT	No extractable text was found; the PDF may be a scanned image.
CONVERSION_FAILED	The conversion engine could not produce a Markdown document.
CONVERSION_UNAVAILABLE	The PDF conversion runtime is not available in this deployment.
UPLOAD_FAILED	The result could not be written to the key-value store.
FILE_TOO_LARGE	The input exceeds the run's memory-derived limit.
TEMP_DIR_FAILED	The run could not create a temporary directory.