Pricing
Pay for pages, not seats.
Start free. Upgrade when you need a larger monthly page allowance.
Create a free account to start with 100 free pages, or sign in to change an existing plan.
or
Create a free account and the first 100 pages are on us, no card required. Paid plans refill your page balance to the plan's monthly amount each billing cycle. Unused pages carry over.
Free
Free
$0 100 free pages to start
The complete product, not a limited preview. Needs an account; does not need a card.
- 100 pages, one-time, yours until used
- 100 MB storage · 50 documents
Starter
Starter
$9 / month
For steady solo work.
- 250 pages per month
- 1 GB storage · 500 documents
Recommended
Pro
Pro
$29 / month
For daily document work.
- 1,000 pages per month
- 5 GB storage · 2,000 documents
Business
Business
$99 / month
Our highest allowance for a single user.
- 4,000 pages per month
- 20 GB storage · 10,000 documents
Prices are in USD and exclude VAT or sales tax. Stripe Tax calculates and adds applicable tax at Checkout.
Every plan includes every feature: AI field extraction, GPU OCR with bounding boxes, review and approval, secure sharing with redaction, the REST API with webhooks and batch, dashboards and exports, team workspaces, custom templates and the multilingual interface. Plans differ only in the page allowance, storage and document count.
One page of a document costs one page of your allowance; an AI redaction scan when sharing costs one page per document. Viewing, exports and dashboards are free.
What counts against your allowance
- A failed analysis is not charged. Pages are metered once extraction has actually finished. If a document fails before that (an unreadable file, a provider outage), nothing is deducted.
- Correcting a field is free. Re-analyzing is not. Editing, confirming and approving extracted values cost nothing. Asking DocAI to read a saved document again runs the full extraction a second time and charges its pages again. It reads up to 20 pages on a re-analysis, against 3 on the first pass.
- Running out stops new analyses, nothing else. Documents you have already saved stay exactly where they are. You can still open, export, share and search them; only new extractions wait for your next monthly refill or for an upgrade.
- Unused pages carry over. A renewal or plan change never clears your balance: each paid month refills it up to your plan's monthly amount, and anything left keeps rolling until you spend it. Pages lapse only when your subscription fully ends. The free starter pages are one-time and always kept.
- Long documents are read in part, not in full. DocAI renders the first 5 pages of a PDF and sends the first 3 to the AI model. That covers invoices, receipts, purchase orders and delivery notes, which is what it is built for. A long contract or a full bank statement will be read only as far as those pages reach.
What makes DocAI different
- Supplier learning, automatically. Every correction you make on a supplier's document is remembered. Next time a document arrives from the same vendor, DocAI injects those hints into the extraction so the same mistake does not recur. Transparent: results show which hints were applied, and you can review or delete them at any time.
- Every field links to its exact source on the page. Click any extracted value to jump to the matching bounding box on the document image. No guessing where the number came from. Bounding boxes are derived from GPU OCR and matched to the extracted text, so you can verify at a glance.
- Flat pricing: every feature, every plan. There are no feature-locked tiers. Free and Business users get identical capabilities: OCR, AI extraction, bounding boxes, review workflow, exports, API, webhooks, workspaces, and templates. Plans differ only in how many pages you can process each month.
- E-invoice XML read first, AI second. Factur-X, ZUGFeRD, KSeF FA(3), and UBL 2.1 invoices carry embedded structured XML. DocAI reads that XML directly (no OCR, no LLM inference cost, no recognition error) and falls back to AI extraction only for documents without a structured data layer.
Questions? hello@synairo.com