Reducto charges per page and per product. Since September 1, 2026, r-1 Parse costs $10 per 1,000 pages, Extract costs $20 per 1,000 pages with parsing included, and Deep Extract costs $40 per 1,000 pages. Jobs sent through the batch queue get 20% off Parse and Extract. For invoice fields that means $0.016 to $0.04 a page before you write any code, plus the engineering to turn JSON into an approved bill in your accounting system. Upload an invoice below to see the finished result without an API key.
PDF, JPG, PNG, BMP, HEIC and TIFF
Drop files here or click to upload
Up to 50 files
Uploading...
Reducto is a strong document API built for engineering teams. The per-page numbers are clear, but an AP buyer still has to work out which product an invoice needs and what sits outside the meter.
r-1 Parse at $10 per 1,000 pages returns layout, text and tables as chunks for a model to read. It does not give you a vendor name, an invoice number or a due date. For typed fields you need Extract, which doubles the rate to $20.
Deep Extract is $40 per 1,000 pages, four times the price of Parse. Scanned bills, multi-page line items and handwritten notes are exactly the pages where a team is tempted to switch it on for everything.
Reducto moved from a single credit meter to per-product rates on September 1, 2026. Reviews, calculators and AI answers written before then still quote the old $0.015 per credit and the 2x agentic multiplier, so budgets built on them are off.
The Zero Data Retention agreement and the Business Associate Agreement are listed on Growth, which is custom priced. A US team handling medical bills or anything under HIPAA cannot stay on self-serve pricing.
Standard includes up to 5 Studio seats and 200 concurrent pages. A month-end backlog from a shared services team can hit the concurrency ceiling, and more seats means a sales conversation.
Reducto returns JSON to an application you build. Upload screens, review queues for low-confidence fields, duplicate checks and the export into QuickBooks, NetSuite or Excel are not part of the per-page price.
These rates come from Reducto's published usage and pricing documentation, which states that the listed prices took effect on September 1, 2026. Confirm them with Reducto before you sign, since the model is new.
The new r-1 parsing model is one cent a page with no feature multipliers. Legacy parsing stays on the list at $15 per 1,000 standard pages and $30 per 1,000 complex pages, with a 2x multiplier for agentic modes.
Extract returns fields against a schema you define and includes the parse step in the price. This is the product an invoice pipeline actually needs, at two cents a page.
The heavier extraction mode for complex or low-quality documents. At four cents a page it costs four times Textract Analyze Expense for the same invoice.
Parse and Extract jobs sent through the batch queue get a 20% usage discount. That brings Extract to $16 per 1,000 pages and Deep Extract to $32 for overnight backfills.
Split is $20 per 1,000 pages, Deep Split $40, Classify $7.50 and Edit $60 ($15 when the page is fully prefilled). Splitting a scanned stack of bills into single invoices is a separate charge on top of Extract.
The pay-as-you-go Standard plan starts with $150 in free usage. At Extract rates that covers about 7,500 invoice pages, enough for a real pilot. Growth and Enterprise are custom priced.
The meter is simple once you pick the right product. The honest comparison needs the build cost too.
Count invoice pages, not invoices, because a three-page bill is three pages. Multiply by $0.02 for Extract, $0.04 for Deep Extract, or 80% of those if the work can wait for the batch queue.
Tip: At 2,000 invoice pages a month, Extract is about $40 and Deep Extract about $80.
If invoices arrive as one scanned PDF, add Split at $20 per 1,000 pages. Then price the developer days to wire the API into review, approval and export, plus the upkeep when a vendor changes its layout.
Put that total next to a flat plan that already includes upload, review, line items and Excel, CSV or JSON export. Upload one of your own invoices above to check the output before you decide.
A document API and an invoice extraction app solve different problems. Here is the honest split.
If you are feeding contracts, filings or manuals into a retrieval pipeline, r-1 Parse at one cent a page is good value and chunked output is exactly what you want.
Reducto offers VPC and on-prem deployment on Enterprise. If documents cannot leave your network, that option is worth the custom contract.
If the goal is vendor bills in a spreadsheet or QuickBooks this week and nobody on the team writes code, an API priced at $0.02 a page still leaves the whole application to build.
A bookkeeping firm processing a few thousand invoice pages a month across clients gets more from a flat plan with line items and exports than from a metered API plus a developer.
Reducto bills per page, and since September 1, 2026 each product has its own rate instead of sharing one credit meter. The table below is the full list from Reducto's usage and pricing documentation. Batch pricing applies only to Parse and Extract.
| Reducto product | What it returns | Per 1,000 pages | Batch queue (20% off) |
|---|---|---|---|
| r-1 Parse | Text, layout and tables as chunks | $10.00 | $8.00 |
| Legacy Parse, standard page | Previous parsing pipeline, 2x in agentic modes | $15.00 | $12.00 |
| Legacy Parse, complex page | Previous pipeline on dense pages, 2x in agentic modes | $30.00 | $24.00 |
| Extract | Fields against your schema, parsing included | $20.00 | $16.00 |
| Deep Extract | Heavier extraction for difficult documents | $40.00 | $32.00 |
| Split / Deep Split | Breaks a combined PDF into documents | $20.00 / $40.00 | No batch discount |
| Classify / Deep Classify | Labels the document type | $7.50 / $15.00 | No batch discount |
| Edit | Fills forms; $15.00 when fully prefilled | $60.00 | No batch discount |
Two add-ons sit on top of Parse: OCR data return at $2 per 1,000 pages, and prompted blocks with custom regions at $5 per 1,000 pages. The advanced chart agent is billed at $0.06 per chart. Spreadsheets are billed by cells, one credit per 1,000 cells in accurate mode and one per 5,000 in fast mode.
Until the end of August, Reducto sold one kind of credit. The rule of thumb in most reviews was $0.015 a credit, one credit for a standard page, two for a complex page, and a 2x multiplier on agentic modes. Reducto's own announcement of r-1 put legacy agentic parsing at 3 to 6 cents a page. The new card replaces that with a flat rate per product and makes r-1 Parse the cheapest way in, at one cent a page all in.
For invoice work the headline cut matters less than it seems. An invoice pipeline pays for Extract, not Parse, and Extract lists at $20 per 1,000 pages. If you built a budget on "about $0.015 a page" from an older review, your real invoice rate is 33% higher on Extract and 167% higher on Deep Extract.
A typical US vendor invoice runs one to two pages. Here is the monthly bill on Reducto's published rates for three common AP volumes, before engineering time.
| Invoice pages per month | Extract | Extract via batch | Deep Extract | Deep Extract via batch |
|---|---|---|---|---|
| 2,000 | $40 | $32 | $80 | $64 |
| 10,000 | $200 | $160 | $400 | $320 |
| 50,000 | $1,000 | $800 | $2,000 | $1,600 |
If your invoices arrive as one scanned stack, add Split at $20 per 1,000 pages to every row. Growth plan volume discounts are negotiated, so above roughly 50,000 pages a month the list rate is a ceiling, not a quote.
The fair comparison is the product that returns invoice fields, not the cheapest parsing meter. Every rate below is the published US list price per 1,000 pages for structured output.
| Service | Product for invoice fields | Per 1,000 pages | Batch or volume discount |
|---|---|---|---|
| Mistral | Document AI (annotations) | $5.00 | Batch API 50% off |
| AWS Textract | Analyze Expense | $10.00 | $8.00 above 1M pages a month |
| Azure Document Intelligence | Prebuilt invoice model | $10.00 | Commitment tiers from $190 a month |
| Reducto | Extract | $20.00 | Batch queue 20% off, Growth volume rates |
| Azure Content Understanding | Document field extraction (West US) | $22.00 | None, consumption only |
| Reducto | Deep Extract | $40.00 | Batch queue 20% off |
On list price Reducto Extract costs twice what AWS and Azure charge for their prebuilt invoice models. You are paying for a schema you control instead of a fixed invoice field list, and for Reducto's handling of dense tables, which is where the hyperscaler invoice models tend to struggle. Reducto says r-1 cut errors by 20% against its own previous pipeline, but that figure is measured against itself, and no public evaluation set was released with it, so test it on your own invoices before you pay the premium.
For an engineering team building a document product, often yes. The schema-driven Extract endpoint, the batch queue and VPC deployment are real advantages, and $20 per 1,000 pages is small next to a developer's salary.
For a finance team that wants vendor bills in Excel or QuickBooks, the API is only the first piece. Reducto returns JSON. Someone still builds the upload flow, the review screen for fields the model is unsure about, duplicate detection, and the export in the column order your ERP expects. That is typically weeks of work before the first invoice lands in the books, plus maintenance every time a vendor redesigns its PDF.
InvoiceExtractor is the finished version of that pipeline. You drag in PDFs, scans or phone photos, the AI reads header fields and every line item, and you download Excel, CSV or JSON. The Starter plan is $49 a month, or $24 a month billed yearly, and covers 2,500 pages on the Base model or 500 on the Pro model. The Pro plan is $149 a month, or $74 billed yearly, for 10,000 Base pages or 2,000 Pro pages. There is no per-product meter, no schema to write and no code.
Pick Reducto when you have developers, the output feeds your own software, and you need to control the schema or run inside your own VPC. Pick a finished invoice extraction tool when the output is a spreadsheet or an accounting import, your team is finance rather than engineering, and you want the first batch processed today rather than after a sprint. If you are not sure, run the same ten invoices through both and compare the line items, not the headline accuracy claim.
Reducto charges per page by product. Since September 1, 2026, r-1 Parse is $10 per 1,000 pages, Extract is $20 per 1,000 pages with parsing included, and Deep Extract is $40 per 1,000 pages. The batch queue takes 20% off Parse and Extract. Standard is pay-as-you-go with $150 of free usage; Growth and Enterprise are custom priced.
One cent a page for r-1 Parse, two cents for Extract and four cents for Deep Extract at list price. Through the batch queue those become 0.8, 1.6 and 3.2 cents. Split costs another two cents a page if you need to separate a combined PDF into individual documents first.
No, but the Standard plan starts with $150 in free usage. That covers about 15,000 pages of r-1 Parse or about 7,500 pages of Extract, which is enough to test on real invoices. After that every page is billed at the published rate.
Only for spreadsheets, which are billed at one credit per 1,000 cells in accurate mode. Documents moved to per-product rates on September 1, 2026. Older reviews that quote $0.015 per credit and a 2x agentic multiplier describe the legacy pricing, which now appears as Legacy Parse at $15 and $30 per 1,000 pages.
No. AWS Textract Analyze Expense returns invoice fields for $10 per 1,000 pages, half of Reducto Extract at $20. Reducto gives you a custom schema and stronger table handling, and its batch queue brings Extract to $16, but on list price Textract is cheaper for standard invoices.
Reducto lists the Business Associate Agreement and the Zero Data Retention agreement on its Growth plan, which is custom priced. The self-serve Standard plan does not list them, so a team processing medical invoices or other protected health information should plan for a sales contract.
Both return fields against a schema you define and include parsing. Deep Extract runs a heavier process meant for difficult documents such as poor scans and long, dense tables, and costs twice as much at $40 per 1,000 pages. Use it for the pages that fail on Extract rather than for every invoice.
Yes. InvoiceExtractor reads invoices and their line items and exports Excel, CSV or JSON with no API key or schema. Plans start at $49 a month, or $24 a month billed yearly, for 2,500 pages on the Base model. You can upload an invoice at the top of this page to check the output first.
AWS Textract pricing per page, Analyze Expense at $10 per 1,000 invoice pages.
Mistral OCR pricing: $4 per 1,000 pages, and Document AI fields at $5.
Azure Document Intelligence pricing, prebuilt invoice model and commitment tiers.
LlamaParse pricing in credits, from 1 to 45 credits a page.
Azure Content Understanding pricing, including the $22 field extraction meter.
Invoice data extraction API that returns invoice fields as JSON.
Invoice line item extraction for every row, not just the totals.
Start turning your invoices into clean, structured spreadsheet data.
USD
per month
per month
$288 charged today
Choose speed vs accuracy when extracting
| Base AI Faster | 2,500 pages |
| Pro AI Best accuracy | 500 pages |
Scale invoice extraction across your whole team with automation.
USD
per month
per month
$888 charged today
Choose speed vs accuracy when extracting
| Base AI Faster | 10,000 pages |
| Pro AI Best accuracy | 2,000 pages |
Enterprise‑grade invoice extraction, security, and controls.
USD
per month
per month
$ charged today
Choose speed vs accuracy when extracting
| Base AI Faster | pages |
| Pro AI Best accuracy | pages |
Secure checkout powered by Stripe
