Pricing: per document, not per page

Most document parsing APIs charge per page. We charge per document. A 1-page doc or a 1,000-page doc — same price, same request.

Parse a 1-page doc or a 1,000-page doc — same price
1 request = 1 document, regardless of page count. No per-page fees, ever.
Office formats are parsed deterministically — no AI model needed, no per-page compute.
100×
cheaper than per-page
PDF parsing APIs
vs
$25–$250
per-page APIs
1,000 docs × 25pg

Tier plans

Free
€0
per document, any page count

Evaluate parsing quality across your document types.

  • 1,000 documents / month
  • 50 AI parses / month
  • 50 daily rate limit
  • 10 MB max file size
  • All 9 output formats
  • 30-day history + replay
  • Community support
Get Free API Key Try in browser, no signup →
Business
€99 / month
€0.000198 per document at full usage

Document-heavy operations with dedicated support.

  • 500,000 documents / month
  • 2,000 AI parses / month
  • 20,000 daily rate limit
  • 50 MB max file size
  • All output formats
  • 365-day history + replay
  • GCS large-file upload (>32 MB)
  • Dedicated support
Start Business
Need something different?

Higher volumes, dedicated infrastructure, on-prem deployment, custom SLAs, private models, or a price that fits your usage shape — we’re flexible and can accommodate almost any need. Plans can be tailored to your workload.

Get in touch →

How pricing works

AILANG Parse charges per document, not per page. One API request parses one entire document — whether it's 1 page or 1,000 pages. The cost is the same. This is fundamentally different from every other document parsing API, which charges per page ($0.001–$0.01/page).

Why? Because we read Office XML directly. A 200-page DOCX is just a zip file of XML — we parse the structure in ~50 ms on a single server instance. No PDF conversion, no OCR, no per-page compute. The cost to us is the same for 1 page or 1,000 pages, so we pass that through.

There are two types of requests:

  • Deterministic requests — Office formats (DOCX, PPTX, XLSX, ODT, ODP, ODS), Email (EML, MBOX), CSV, Markdown, HTML, EPUB. Instant, exact, no AI model involved.
  • AI parses — PDFs and images. These invoke a multimodal AI model. Metered separately because the underlying compute cost is significantly higher.

Full tier comparison

Every tier includes all 14 input formats. 1 request = 1 document, any page count.

Feature Free Pro €29/mo Business €99/mo
Documents / month 1,000 100,000 500,000
AI parses / month 50 500 2,000
Daily rate limit 50 5,000 20,000
Max file size 10 MB 25 MB 50 MB
Per-page fees None None None
Cost per DOCX (25 pages avg) ~€0 €0.00029 €0.000198
Per-page API cost (same doc) $0.025–$0.25 per doc at typical $0.001–$0.01/page rates
Output formats All 9 formats All 9 formats All 9 formats
History retention 30 days 180 days 365 days
Request replay Yes Yes Yes
Unstructured API compat Yes Yes Yes
File upload (all tiers) Yes Yes Yes
GCS large-file upload (>32 MB) Yes
Support Community Email Dedicated

AI parse file size limits

AI-powered parsing (PDF and images) has per-format limits by tier. Deterministic formats (Office, CSV, etc.) are bounded only by the max file size above.

Format Free Pro Business
PDF 10 MB 25 MB 50 MB
Images (PNG, JPG) 10 MB 25 MB 50 MB

Video and audio parsing is available via CLI and WASM with your own AI key. See Self-Host.

Per-document vs per-page: how to compare

Most document parsing APIs charge per page. When comparing, multiply their per-page rate by your average page count. Here's the typical range for 1,000 documents at 25 pages each (25,000 pages):

AILANG Parse (Pro)
~100× cheaper
per document — any page count, no per-page fees
93.9% composite on OfficeDocBench. Deterministic Office parsing. Track changes, merged cells, comments preserved.
Per-page APIs (typical range)
$25–$250
$0.001–$0.01/page × 25,000 pages
Costs scale linearly with page count. A 100-page doc costs 100× more than a 1-page doc. Check your provider's current pricing.
Cloud OCR services
$37–$250
$0.0015–$0.01/page × 25,000 pages
Require PDF or image input. Convert Office docs to PDF first, losing structured data. Check cloud provider pricing pages.

Key difference: per-page costs scale linearly with document length. A 100-page document costs 100× more than a 1-page document. With AILANG Parse, both cost one request. The longer your documents, the bigger your savings. Check competitors' current pricing pages to verify — rates change.

Why per-document pricing matters

You're not paying for document parsing

You're paying for AI systems that don't hallucinate when grounded in documents. Better parsing = fewer hallucinations = cheaper total system. At 93.9% composite (and 100% format coverage) vs 27–68% coverage-adjusted from alternatives, every €29/month saves hours of debugging bad RAG outputs.

Per-page pricing penalizes exactly the documents that matter most — long contracts, detailed reports, comprehensive manuals. With per-document pricing:

  • Your costs are predictable — they scale with document count, not page count
  • Long documents are not penalized — a 500-page contract costs the same as a 1-page memo
  • You can send the original Office file instead of converting to PDF — better quality, lower cost, no AI quota used

Frequently Asked Questions

Is pricing really per document, not per page?

Yes. One request = one document, regardless of page count. A 1-page DOCX and a 1,000-page DOCX each consume one request from your monthly quota. There are no per-page fees, no page-count multipliers, no hidden charges based on document length. This applies to all formats — Office, PDF, images, everything.

When should I use AILANG Parse?

When your AI system ingests Office documents (DOCX, PPTX, XLSX, ODT) and you need structured, accurate output. We read Office XML directly — no PDF conversion, no OCR, no information loss. Track changes, merged table cells, comments, and formatting are all preserved. If your RAG pipeline or agent system processes business documents, this is built for you.

When should I NOT use AILANG Parse?

If you only parse scanned PDFs or handwritten documents, a dedicated OCR service may be more appropriate. Our AI parsing handles PDFs, but our structural advantage is in Office formats where we bypass the PDF-then-OCR pipeline entirely. For pure scan workloads, we add an AI layer but don't offer a unique cost or quality advantage.

Why is it so much cheaper than competitors?

Competitors convert your DOCX to PDF, render it as images, then run OCR or vision models to recover the text — expensive compute for every page. We read Office XML directly. A 200-page DOCX takes ~50ms to parse deterministically. No AI model needed, no per-page cost. The cost to us is the same for 1 page or 1,000 pages.

What are “AI parses” vs regular requests?

Regular requests use our deterministic parser — fast, cheap, and reliable for Office formats, CSV, Markdown, HTML, EPUB. No AI model involved. AI parses invoke a multimodal AI model for PDFs and images. Metered separately because AI inference costs significantly more than deterministic parsing.

What does “daily rate limit” mean?

A burst-protection cap — the maximum documents you can parse in one calendar day. Prevents runaway scripts from exhausting your monthly quota in one burst. The monthly limit is the binding constraint for most users.

Do you support the Unstructured API format?

Yes. Drop-in compatible endpoint at /general/v0/general. Change the base URL and get higher accuracy at a fraction of the cost. All tiers.

What output formats are available?

All tiers: JSON (Block ADT with 9 variants), Markdown, plain text, HTML, Quarto (.qmd), A2UI (agent-to-UI protocol), and Unstructured-compatible output. No output formats are gated by tier.

Can I self-host?

Yes. Run locally via Docker or CLI with unlimited deterministic parsing, no API key needed.

Need higher limits or custom pricing?

Contact docparse@sunholo.com for enterprise volume, SLAs, or dedicated deployments.

Get a free API key in 30 seconds. No credit card. 1,000 documents/month — any page count.