How does the $0.0001 price compare with Textract Signatures?

On the Amazon Textract pricing page, Analyze Document with Signatures is $0.0035 per page for the first 1 million pages in a month and $0.0014 per page after that, in the US West (Oregon) examples. For 100,000 pages their example total is $350. The same count of single-page images on Signature Detector is 100,000 credits, which is $10. Textract's figure is per page and includes OCR text. Signature Detector's figure is per detection request and returns signature boxes only.

What does Google Document AI return for a signature?

Google's custom extractor signature type returns a derived entity. When a signature is found, the sample response has mention text Detected and normalized signature_value true. When it is not found, the entity is omitted. Google's documentation says textAnchor and pageAnchor are not populated for that signature entity, so the field does not include a bounding box. Custom extractor list price is $30 per 1,000 pages for the first 1,000,000 pages in a month, then $20 per 1,000.

What does Azure custom extraction cost?

Azure's East US retail meter S0 Custom Pages is $30 per 1,000 pages, effective 1 June 2024. Custom generative extraction and batch custom extraction are the same $30 per 1,000 on that price list. Commitment tiers are lower if you prepay a monthly page volume. Microsoft announced the cut from $50 to $30 per 1,000 pages on 1 June 2024. Custom extraction requires a model you train. It is a different meter from Azure's prebuilt models.

Is this an accuracy comparison?

No. These pages cite published prices and documented output fields. Signature Detector does not publish an accuracy number against Textract, Azure, or Google.

When is Textract still the better fit?

When the same call also needs the page's words, forms, tables, or queries. Textract Signatures is one feature of Analyze Document, and the pricing page says the response includes OCR. Signature Detector does not return the text of the document.

Textract Signature Alternative

Prices, in the units each vendor publishes

One uncached image is one credit. A PDF is one credit per page, so a 100,000-page corpus is $10 either way when none of the files are cached. A repeated identical PDF is not charged again (creditsUsed is 0). See PDF signature detection.

Service	Published price	100,000 pages or images	Signature output
Signature Detector	$10 / 100,000 detections ($0.0001 each)	$10 for 100,000 uncached images or PDF pages	Pixel box (`x1`, `y1`, `x2`, `y2`) and a confidence score
Amazon Textract, Analyze Document Signatures	$0.0035 / page for the first 1M pages, then $0.0014 / page (US West (Oregon) examples)	$350 at the first-tier rate	`SIGNATURE` blocks with geometry, plus OCR text. Confidence on Textract blocks is documented as 0–100.
Azure Document Intelligence, custom extraction	$30 / 1,000 pages (S0 Custom Pages, East US)	$3,000	Whatever fields you labeled when you trained the custom model
Google Document AI, custom extractor	$30 / 1,000 pages for the first 1,000,000 pages, then $20 / 1,000	$3,000 at the first-tier rate	Signature entity: `signature_value: true` and text "Detected", or the entity is omitted. No `pageAnchor` on that field.

Worked totals at the list prices above:

1,000,000 single-page images. Signature Detector $100. Textract Signatures $3,500. Azure S0 custom extraction $30,000. Google custom extractor $30,000 (the $20 tier starts after 1,000,000 pages).
5,000,000 pages. Signature Detector $500 if each page is one uncached image request. Textract Signatures $9,100, which is the total in Amazon's pricing example 9 ($3,500 for the first million and $5,600 for the next four million). Google custom extractor $110,000 ($30,000 for the first million and $80,000 for the next four million at $20 per 1,000). Azure S0 custom extraction $150,000, because that meter has no automatic volume break.

What the Textract price includes that this API does not

Amazon's pricing page describes Analyze Document Signatures as detection of handwritten signatures, electronic signatures, and initials, and says the feature also uses OCR to extract text. FeatureTypes includes SIGNATURES, and a SIGNATURE block carries location and confidence. Geometry on Textract blocks is a bounding box and a polygon, with box values as ratios of the page, not raw pixels.

Adding Forms, Tables, or Queries is priced separately and stacks. The $0.0035 figure is the signatures example, not a bundle. New AWS accounts can analyze 1,000 pages per month with Signatures only, for three months, under the free tier described on that page. Prices in the examples are US West (Oregon). Another region can differ.

One sentence on Amazon's example 8 multiplies 100,000 pages by $0.00035 and still prints a total of $350. $350 divided by 100,000 is $0.0035, and example 9 uses $0.0035 for the first million pages. This page uses $0.0035 and $0.0014.

Google's signature field is a yes, without a box

Custom extractor list prices are on the Document AI pricing page. The signature behavior is documented on Derived field and signature detection. You define a schema entity with data type signature and method Derived. Supported foundation-model versions are listed on that page, and the derived-fields section is marked preview. Hosting a custom processor version is an extra $0.05 per hour on the pricing page, which is $438 for one version deployed all year. Those hosting dollars are not in the table above.

If you need the mark's rectangle, a yes/no entity does not give it to you. Signature Detector's response is the rectangle and a score, and nothing else: no OCR line items, no schema to label.

Azure custom models

The Azure Document Intelligence pricing page lists custom extraction per 1,000 pages. The dollar amount checked for this page is the East US retail meter S0 Custom Pages: $30 per 1,000, effective 1 June 2024. The same list shows S0 Custom Generative Pages and S0 Batch Custom Extraction Pages at $30 per 1,000. Microsoft's 1 June 2024 announcement states the custom extraction price moved from $50 to $30 per 1,000 pages. Commitment tiers on the retail list are cheaper. For example, the East US "Commitment Tier Custom Azure 100K" unit is $2,400 per month for 100,000 pages, with overage at $24 per 1,000. Training a custom neural model is free for the first 10 hours and then $3 per hour, according to the FAQ on the pricing page. Prebuilt models, including the contract model named on that page, are a separate meter. This comparison uses custom extraction because that is the published ~$30 per 1,000 pages rate.

What you call here

One page image

curl -X POST https://signature-detector.com/api/detect \
  -H "Content-Type: application/json" \
  -H "x-api-key: your_api_key_here" \
  -d '{
    "imageUrl": "https://example.com/contract-page.jpg"
  }'

The response fields are documented on the signature detection API page. There is no training step and no cloud project. The confidence value is the detector's score. The docs sample uses 0.192 against a fixed cutoff of 0.1. That scale is not Textract's 0–100 Confidence, so do not compare the two numbers directly.

Our price, from the product

$10 per 100,000 credits, which is $0.0001 each. An image that is not served from cache spends 1 credit. A PDF spends 1 credit per page.
10 credits are added when the account is created.
$1 minimum purchase. That is 10,000 credits. There is no subscription.
Credits do not expire. The balance stays on the account until a detection spends it. A failed detection puts the credit back. Repeating an image or an identical PDF returns cached: true and does not spend a credit. On a PDF that hit is creditsUsed: 0.
Buy credits from the dashboard or estimate a volume on the pricing page.

The arithmetic, with the same citations, is also written up as Textract vs Signature Detector: cost comparison. Sources: Amazon Textract pricing, Textract Block, AnalyzeDocument, Document AI pricing, Google signature detection, and Azure Document Intelligence pricing. None of these prices include tax, support plans, or data transfer.

Textract signature alternative