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.
Evaluate parsing quality across your document types.
Ship to production. Advanced outputs for RAG and agents.
Document-heavy operations with dedicated support.
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.
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:
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 | Dedicated | |
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 |
|---|---|---|---|
| 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.
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):
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.
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:
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 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.
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.
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.
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.
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.
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.
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.
Yes. Run locally via Docker or CLI with unlimited deterministic parsing, no API key needed.
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.