Amazon Textract
AWS API that reads text, forms, tables and signatures from uploaded documents.
Opens the vendor's own site.
Jobs covered
Open a task to see exactly what the vendor documentation supports.
Extract
Browse all extract toolsdetecting typed and handwritten text and extracting text, forms and tables, and extracting specified information with queries, through the Amazon Textract Document Analysis and Analyze Document APIs on image and PDF files
OCR
Browse all ocr toolsdetecting typed and handwritten text in images and PDFs through Amazon Textract text detection, with synchronous processing for single page documents and asynchronous operations for multipage documents
Where your files are processed
Reported to upload your files
The vendor states that files are uploaded to their servers. We have not independently tested this.
Documents are uploaded to AWS: you send bytes or an S3 object to a regional Textract endpoint and AWS processes them in that region. There is no on-premises or self-hosted Textract edition.
What it does
A managed AWS service that runs OCR and document-structure analysis on scans and PDFs you send to it. As well as plain text and handwriting it returns form key-value pairs, table cells, page layout, detected signatures, invoice and receipt line items, identity-document fields and lending classifications, all as JSON with confidence scores and coordinates.
Single pages can be processed synchronously; multi-page jobs run asynchronously out of Amazon S3. Nothing is installed locally and there is no on-premises edition.
Best for
Developers already on AWS who want per-page, machine-readable extraction of forms, tables, invoices and IDs without running infrastructure.
Worth knowing
The limits depend on the mode rather than on the service. Synchronous input is 10 MB, and a PDF or TIFF is one page; asynchronous PDF or TIFF is 500 MB and up to 3,000 pages, while JPEG and PNG stay at 10 MB.
There is no ongoing free plan: new customers get a three-month introductory allowance that varies by API and excludes Custom Queries, and 100 pages a month on some APIs, so it is a trial rather than a tier. Custom Queries and the Forms and Tables APIs have their own price bands, and AWS storage and notification charges sit alongside the API cost.
Detection of a signature is not authentication, and a detected checkmark is not a tested one.
Details, and where each comes from
Vendor statements, our own reading of their pages, and the things nobody has established yet, each answered separately. Where we have not confirmed something it says so rather than guessing, and a row nobody has established is greyed out.
- Versions covered
- Web · API
- Website
- aws.amazon.com
- Requires an account
- Yes
- Free tier
- No
- Watermark on trial output
- No
- Maximum file size
- 500 MB
- Batch processing
- Yes
- Data residency
- Multiple regions
- Open source
- No
- Self-hostable
- No
- Licence
- Not verified
- Redaction
- No redaction feature
- Programmatic access
- HTTP API
- Last availability check
- Responded in 190 ms 21 October 2026
Checked automatically by our own runner, not reported by the vendor. 100% of our last 30 checks got a response.
What the redaction verdict rests on
From the vendor's own documentation, read 22 September 2026. Not tested by us.
What that does not establish
The vendor’s own site was read on 22 September 2026 and contains no mention of redaction, sanitisation, blackout or whiteout. Absence from a vendor’s own material is not proof that a feature does not exist - it may be undocumented, or on a page the reading did not reach - but for a NEGATIVE claim it is the strongest evidence available, and it is weaker than a positive quote.
Recorded as such rather than presented as certainty.
What the vendor says
Claims are labelled rather than repeated as fact. Where we have checked something ourselves, it says so and shows when. Checked means we read the vendor's own page and it says what we recorded - not that we tested the behaviour, unless the note says we did.
-
“With Amazon Textract, you only pay for the documents you analyze. There are no minimum fees or upfront commitments.”
Checked by us - the source says this Read 22 September 2026
Confirmed verbatim: the sentence appears unaltered under a bullet labelled "Low cost". The page's own continuation is worth recording because it is the practical half - "You can get started for free, and save more as you grow with our tiered pricing model" - and the free tier and the tiers themselves are on a separate pricing page that this one does not state.
Note the contrast with Apryse's pricing claim, verified earlier in this campaign, which quoted an entry figure with no billing period: this claim quotes no figure at all and is the more honest of the two for it. No API call was made and nothing was billed.
Standing: Re-read and matched byte for byte
Check it yourself: the source, the SHA-256, and how we checked
- From
- the text a browser renders
- SHA-256
- b3cc0cfcd8e0446f5271ffd8ba7453321965ee2a316cc45974b9f1e6b69c62c7
How we checked
Re-verified verbatim on 2026-09-22: https://docs.aws.amazon.com/textract/latest/dg/what-is.html re-rendered in headless Chrome with scripts/render-page.mjs and the claim's wording found in the rendered text. Originally verified on 2026-09-21. sourceHash is the SHA-256 of the re-rendered text, held as rv-docs-aws-amazon-com-textract-latest-dg-what-is-html-0a72d657.txt.
-
“For synchronous operations, JPEG, PNG, PDF, and TIFF files have a limit of 10 MB in memory. PDF and TIFF files also have a limit of 1 page. For asynchronous operations, JPEG and PNG files have a limit of 10 MB in memory. PDF and TIFF files have a limit of 500 MB in memory and a limit of 3,000 pages.”
Checked by us - the source says this Read 22 September 2026
Confirmed verbatim: every figure in the claim appears unaltered in one paragraph of the limits page. This is the most precisely stated claim in the tier and it is worth noting why - the limits are asymmetric between the two APIs and a reader who assumed one set of numbers applies to both would be wrong in both directions.
Synchronous operations cap at 10 MB and a single page for PDF and TIFF; asynchronous operations allow 500 MB and 3,000 pages for the same formats, but still cap JPEG and PNG at 10 MB. Recorded as the vendor's own documentation; no request was made to the service and no control was run.
Standing: Re-read and matched byte for byte
Check it yourself: the source, the SHA-256, and how we checked
- From
- the text a browser renders
- SHA-256
- 83c3e80903f7c04cca98f2f36091f8c4cea00b575dacd45bc676018f78d6c02a
How we checked
Re-verified verbatim on 2026-09-22: https://docs.aws.amazon.com/textract/latest/dg/limits-document.html re-rendered in headless Chrome with scripts/render-page.mjs and the claim's wording found in the rendered text. Originally verified on 2026-09-21. sourceHash is the SHA-256 of the re-rendered text, held as rv-docs-aws-amazon-com-textract-latest-dg-limits-document-htm-93dfd4ec.txt.
Each claim is quoted from the vendor's own pages. This section describes what we have and have not checked - it is not a judgement about anyone.
Sponsorship is distribution, not an endorsement or a promise of search rankings, AI citations, traffic, or sales. This listing is a free entry, compiled from published sources.