PDF Procurement Document Extractor

Find and extract procurement-related PDFs from company websites — certificates, reports, datasheets, price lists.

Standalone Apify actor for supplier shortlist prep across CEE/EU. Pairs well with the Google Maps CEE Supplier Shortlist Builder and Website Intelligence + Contact Extractor when you already have company websites.

Optional next step after Maps / Website Intel: feed website URLs into this actor to surface procurement PDFs (ISO certs, reports, datasheets).

What it does

• Crawls a company site or downloads/investor page for PDF links.
• Follows download-like paths (/downloads, /dokumentumok, /publications, /investor, /quality, …).
• Downloads each PDF (capped) and extracts text (no OCR at launch).
• Classifies documents: certificate, assurance/annual report, datasheet, price list, catalog, policy.
• Detects ISO hints (9001, 14001, 45001, 27001, …) from filename + text.
• Skips full text extract for oversized PDFs (metadata still saved) so Store QA stays fast.
• Built for n8n, Make, Zapier, and vibe-coder agent workflows.

Who it's for

1. Procurement teams collecting supplier certificates and reports before outreach.
2. Sales / channel teams mining price lists and catalogs from brand sites.
3. Agencies running document discovery for client shortlists.
4. Vibe coders wiring website → PDF evidence into CRM / n8n / Make / Zapier.
5. Market researchers auditing public compliance document footprints.

What this actor is — and is not

This actor helps you discover and extract public PDF evidence. It does not verify certificate authenticity, accreditation status, or document legal validity.

Use the output for:

• supplier document discovery
• shortlist preparation
• compliance evidence collection

Do not use it as the sole basis for vendor qualification or contract award.

What data you get

Field	Description
sourceWebsite	Input website / start URL
pdfUrl, fileName, documentTitle	Document identity
documentType	certificate, assurance_report, annual_report, datasheet, price_list, catalog, policy, other
isCertificate, isPriceList, isDatasheet, isAnnualReport	Boolean flags
isoHints	Detected ISO labels when present
pageCount, byteSize	PDF metadata when extracted
textExcerpt	Truncated extracted text
foundOnPage	HTML page where the link was discovered
scrapedAt, error	Provenance / skip or failure notes

How to use it

1. Configure input — website or downloads URL, max PDFs, extract text, proxy.
2. Run on Apify cloud (recommended) or locally with apify run.
3. Download dataset as JSON, CSV, or Excel from the run page.

Input presets

MOL Group annual & assurance PDFs (Store / QA example)

{
  "websites": ["https://www.molgroup.info/en/investor-relations/annual-reports"],
  "maxDocuments": 3,
  "maxPagesPerSite": 3,
  "extractText": true,
  "maxTextChars": 3000,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Store input prefill uses this small cap so Apify automated QA finishes within 5 minutes. For production, raise maxDocuments (e.g. 20–100).

Published Store example: MOL Group annual & assurance PDFs — procurement docs.

After Google Maps / Website Intel

{
  "websites": [
    "https://example-supplier.hu/downloads",
    "https://example-manufacturer.ro/dokumentumok"
  ],
  "maxDocuments": 20,
  "maxPagesPerSite": 8,
  "extractText": true,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Link discovery only (no text extract)

{
  "websites": ["https://www.molgroup.info/en/investor-relations/annual-reports"],
  "maxDocuments": 10,
  "extractText": false,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Output example

{
  "sourceWebsite": "https://www.molgroup.info/en/investor-relations/annual-reports",
  "pdfUrl": "https://molgroup.info/en/show/publications/annual_reports/2025/mol-csrd-limited-assurance-report-eng-2025.pdf",
  "fileName": "mol-csrd-limited-assurance-report-eng-2025.pdf",
  "documentTitle": "mol csrd limited assurance report eng 2025",
  "documentType": "assurance_report",
  "isCertificate": false,
  "isPriceList": false,
  "isDatasheet": false,
  "isAnnualReport": false,
  "isoHints": [],
  "pageCount": 5,
  "byteSize": 387120,
  "textExcerpt": "Limited assurance …",
  "foundOnPage": "https://www.molgroup.info/en/investor-relations/annual-reports",
  "scrapedAt": "2026-07-20T09:30:00.000Z",
  "error": null
}

Proxy, size limits, and blocked sites

Feature	Support
Apify Proxy RESIDENTIAL	Recommended for bot-walled corporate sites
Browser fallback	Used when HTML HTTP fails
Oversized PDFs	Metadata saved; text extract skipped above maxPdfBytes (default 2.5 MB)
No OCR	Scanned image-only PDFs may return empty textExcerpt

Tips to avoid Store QA timeouts

• Keep prefill / public-task demos at maxDocuments: 3.
• Prefer a downloads/investor page with a few mid-size PDFs over a mega catalog portal.
• Raise caps only for paid production runs.

Secrets and local development

• Never commit apify/pdf-procurement-document-extractor/.env — it is listed in .gitignore.
• apify push excludes .env via .gitignore and .actorignore.
• Docker builds exclude .env via .dockerignore.
• On Apify cloud, customers configure Proxy via input — not from your local .env.
• For local runs only, copy .env.example → .env and add APIFY_TOKEN if you call the platform.
• Only .env.example (placeholders, no real keys) is in git.

FAQ

How is this different from a generic PDF-to-text actor?
This actor discovers PDFs on company sites and adds procurement classification flags (certs, reports, datasheets, ISO hints) — not only raw text dump.

Does it work outside CEE?
Yes. Marketing ICP is CEE/Europe; any public website with PDF links can work.

Do I need an LLM key?
No. Classification is deterministic keyword/heuristic. No Apify AI event fee.

Why was text skipped on a large annual report?
maxPdfBytes protects run time. Metadata is still saved; raise the limit for production if you need full text.

How many PDFs can I get per run?
Up to 200 (maxDocuments). Prefill/demo stays at 3.

Local development

cd apify/pdf-procurement-document-extractor
npm install
npm test
npm run build
npm run dry-run
apify run --input-file=INPUT.smoke.json

Dry run defaults to MOL Group annual reports with 3 PDFs. Install Playwright Chromium once for local browser fallback: npx playwright install chromium.

Pricing

Pay-per-event: $5.00 per 1,000 document records (document event @ $0.005), minimum charge $0.10 per run.

Enable Pay per event + usage in Apify Console so platform compute/proxy costs are passed through to users.

Related actors (suite)

Actor	Role
Google Maps CEE Supplier Shortlist	Discover local suppliers
Website Intelligence + Contact	Enrich websites (contacts, tech, compliance)
Dealer / Distributor Locator	Expand brand locator pages
PDF Procurement Document Extractor (this Actor)	Find certs / reports / datasheets
Website Change Monitor	Watch pages for changes

Each Actor runs standalone; use them as an optional suite when chaining discovery → enrichment → monitoring.

Feedback

Open an issue on the actor’s Apify page or contact Vibe Coder's Life.