Start free with a monthly usage allowance, move to the Team plan when you're running in production, or run Datalab in your own infrastructure. Every plan draws on the same per-processor rates — you pay for the processors each request runs.
TRUSTED BY
PLANS
Compare plans.
Start free, scale on Team, or run Datalab in your own infrastructure. Every plan runs the full hosted API at the same per-processor
rates.
If your company is under three years old, has raised less than $15M, is under $5M in ARR, and document processing is core to what
you are building, our startup program takes a third off both the per-page rates and the Team fee for twelve months.
Each processor has its own flat rate per 1,000 pages.
Processors
Processor
What it does
Rate (per 1,000 pages)
Convert — lite
Fast CPU OCR to markdown or HTML: text, tables and lists; no images or LLM passes
0.375
Convert — fast
Document to markdown or HTML, one model pass
1
Convert — balanced
Document to markdown or HTML, checked against the PDF text layer
3
Convert — accurate
Highest-fidelity conversion
7.50
Segment — page level
Detect boundaries between combined documents at the page level. Returns the page ranges and titles only — no OCR text
0.375
Segment — block level
Boundaries can fall mid-page at specific content blocks, driven by an optional custom prompt ("start a new segment at each invoice"). Returns the parsed markdown, so it also runs Convert in the mode you pick and bills its rate
Convert + 0.375
Extraction — turbo / fast
Structured fields, single pass. Turbo returns JSON only; fast adds per-field citations and runs on Convert
4.50
Extraction — balanced
Structured fields, multi-pass with per-field verification
11.25 + fees*
Extraction — accurate
Structured fields, multi-pass with per-field verification, strongest model
15 + fees*
Custom processor
Your own pipeline
15
Agent processor
Document agent workflows — the successor to custom processors, for higher-accuracy and more involved document work
7.50‡
Eval
Compare 2 to 4 versions of a pipeline on your documents, page by page or field by field
15
Form fill
Populate a form's fields
7.50‡
Redaction
Remove information from a document by a redaction spec
7.50‡
Form fields
Find a form's fillable fields and their labels
3
Add-ons
Add-on
What it does
Rate (per 1,000 pages)
Track changes
Diff revisions in a Word document
+4.50
Chart understanding (add-on)
Adds chart parsing to a run
+2.25
Infographic (add-on)
Adds infographic parsing to a run
+3
Cross-page merging (beta)
Stitch content split across a page break back together: tables (long or wide), paragraphs broken mid-sentence, and lists
Variable**
Word bounding boxes
Identify bounding boxes at the word level with confidence scores so you can efficiently route OCR issues for human review and audit mistakes in your pipeline
+1.50
Table cell and list bounding boxes
Identify a bounding box for each table cell and each list item, instead of one box for the whole table or list. Free with word bounding boxes, which they turn on.
Free
* Around 5% of Balanced and Accurate extractions bill above our fixed per-page rates. That happens when the extraction agent has to
work through a document — long documents, and schemas that pull many repeated rows, are the usual cases; the rate covers the rest, and
you are never charged both the rate and the full compute. When it does apply it is typically a dollar or two, more for a long or
unusually dense document (Accurate runs a stronger model, so it both includes and can bill more). Schemas over 750 fields aren't
supported.
** Cross-page merging is in beta. It has no per-page rate — it bills only a variable compute surcharge reflecting the merge work your
document actually needs, which is typically a few cents per document.
‡ Form filling, redaction and agent processors run an agent over your document, so their cost follows the work it needs rather than
its page count. Almost every document bills this rate and nothing more; an unusually demanding one bills what the work actually cost
instead. The rate is a floor, not something a surcharge adds to, so you are never charged twice for the same run.
Spreadsheets bill by cells, not pages — 2,500 cells per page capped at $0.60 per sheet (simple) or 500 cells per page (intelligent),
chosen automatically. Processing in the EU adds 25% to the summed total.
Sharing data for model training takes 25% off the per-page rates above. It does not cover the fees (* and **), requests
processed in the EU, or the startup program's discount (one discount applies). Turn sharing on or off at any time in your data
retention settings.
ESTIMATE YOUR COST
Test out usage scenarios.
Pick the processors a typical request runs and your monthly volume, then get an estimate. It sums each processor exactly the way a
real request does, starting with the data-sharing discount. It is an estimate, not your final bill.
10,000pages / month
01M
PROCESSORSPRICE PER 1,000 PAGES
1.33
PDFs, slides, and more to markdown, HTML, or JSON
0.50
Split a combined file into page-ranged documents. Add Convert for block-level boundaries and parsed markdown
15 + fees*
Find key information with extraction schemas
20
Shape output with your own natural-language rules
20
Compare 2 to 4 versions of your pipeline, page by page or field by field. Each version's run bills at its own rate
10 + fees*
Populate a form's fields
10 + fees*
Document agent workflows for more involved document work
4
Find a form's fillable fields and their labels
ADD-ONS
+6
Diff a Word document's revisions
+3
Add chart parsing to a run
+4
Add infographic parsing to a run
+2
A box and confidence score for every word
Free with word bboxes
Per-cell table bounding boxes (includes word bboxes)
Free with word bboxes
Per-item list bounding boxes (includes word bboxes)
* Around 5% of Balanced and Accurate extractions bill above their fixed per-page rate, and Form fill can too. It applies when the
agent has to work through a document — long documents, and schemas that pull many repeated rows. When it does, it is typically a
dollar or two, and it isn't included in this estimate.
Spreadsheets bill by cells, not pages — 2,500 cells per page capped at $0.60 per sheet (simple) or 500 cells per page
(intelligent), chosen automatically.
FAQ
Common questions.
What counts as a page?
Rates are per 1,000 pages. Spreadsheets count pages by cells: the note under the rate card gives the counts. The page count for a request appears in its breakdown.
What does Team include for the $400 fee?
Team includes $400 of usage a month against the same rate card, then meters anything beyond it.
How is extraction priced for large or dense schemas?
Around 5% of Balanced and Accurate extractions bill above our fixed per-page rates; the rest pay exactly the listed rate. Those rates already include an allowance for the extraction agent, and only a document that needs more compute than its rate covers bills the difference — so you are never charged both the rate and the full compute for the same document. It applies when the agent has to work through a document: long documents, and schemas that pull many repeated rows, are the usual cases. When it does apply it is typically a dollar or two, more for a long or unusually dense document (Accurate runs a stronger model, so it both includes and can bill more). Schemas with more than 750 input fields are not supported — split the extraction into multiple requests.
How is form filling priced?
Form filling is $10 per 1,000 pages, and almost every form bills exactly that. It runs an agent over your form, so its cost follows the work the form needs rather than how many pages it has — a dense one-page form can take more work than a long simple one. An unusually dense form bills what the work actually cost instead of the rate; the rate is a floor rather than something a surcharge adds to, so you are never charged twice for the same fill. A failed fill bills nothing.
How is redaction priced?
Redaction starts at $10 per 1,000 pages on the current rate card. A run bills the greater of its page rate and its compute charge, never their sum. Completed runs are charged even when verification withholds the PDF. The report contains counts and check identifiers, without removed text. Existing contract and legacy plan rates may differ; the playground quotes your team's rate.
How is cross-page merging priced?
Cross-page merging is in beta. It carries no per-page rate of its own — it is an option on a parse, and adds a variable compute charge for the merge work your document actually needs on top of what the parse costs. That charge is typically a few cents per document, and a cached result bills nothing extra.
How does the monthly allowance work?
Both Free and Team include a dollar-denominated monthly allowance. Usage draws it down first, and anything beyond bills at the rates above. It resets each cycle and does not roll over.
Do I need a card on the Free plan?
Not to use your monthly allowance. To continue past it, add a card, which turns on pay-as-you-go at the rates above.
Do you offer discounts?
Yes. Our startup program offers discounted API rates for early-stage companies. We also do annual terms for customers with committed use, and can include high-volume discounts based on the volume you're anticipating. Talk to sales →
Who needs to pay for the models?
Our models are open source. The weights are free for research and personal use, and free for commercial use if your company is under the threshold on both revenue and funding: $2M for Chandra and $5M for Marker and Surya. Above either threshold, commercial use needs a license. We have a startup program with discounts on our API for larger companies, and our sales team can work with you on high-volume discounts or an on-prem license based on your needs.
Do you train custom models?
Yes. Our research team can work with you to understand the edge cases you're running into — get in touch. Contact sales →