Azure AI Document Intelligence
Microsoft's cloud OCR and document extraction API, with an on-premises container option.
Opens the vendor's own site.
Jobs covered
Open a task to see exactly what the vendor documentation supports.
Extract
Browse all extract toolsextracting text, tables, document structure and key-value fields from documents with the cloud-based Azure AI Document Intelligence service, using the Read, Layout, prebuilt or custom models, with fields returned as strongly typed values, through the REST API, C#, Python, Java and JavaScript SDKs or Document Intelligence Studio
OCR
Browse all ocr toolsextracting printed and handwritten text from PDF, JPEG, PNG, BMP, TIFF and HEIF inputs with the Read and Layout models of the cloud Azure AI Document Intelligence service, with optional add-ons for high resolution extraction, font properties, formulas, barcodes and language detection
Convert
Browse all convert toolsconverting an analog or image-only PDF into a PDF with an embedded searchable text layer through the Searchable PDF capability of the cloud Azure AI Document Intelligence service, using the prebuilt-read model with the pdf output option
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. This covers the web service. We have not checked desktop app processing or self-hosted instance, and we do not extend the finding to them.
Documents are sent to an Azure endpoint; the service stores them encrypted in the same region as the resource and deletes input and results 24 hours after analysis. Separately, Microsoft ships Linux Docker containers so Read and Layout can run on your own host - but connected containers still need an Azure key for billing, and disconnected ones need a purchased commitment plan and an expiring licence file.
What it does
A cloud service that converts documents into structured data. It runs OCR and layout analysis over PDFs, scans, images and Office files, and layers prebuilt or customer-trained models on top for invoices, receipts, IDs, bank statements and tax forms.
Results come back as JSON, and the Read model can optionally emit a searchable PDF with an embedded text layer. Unusually for this category, the same models can also run in Docker containers on your own hardware, including fully disconnected environments.
Best for
Teams that need high-volume field extraction with confidence scores, especially organisations already on Azure that want regional processing and an on-premises escape hatch.
Worth knowing
The 500 MB field belongs to a tier, an API version and a model rather than to the service as a whole, so it is not recorded as a product-wide maximum. Model coverage depends on the version: the v4.0 container installation documentation lists Read and Layout, with additional models on v3.1 and v3.0, so the container offering does not mirror the cloud model set at every version.
Containers run on a supported Docker host, which does not have to be a Linux operating system itself, and the supported hosts are documented separately. Connected containers send billing traffic to Azure, which is not document upload, and disconnected licensing is a separate arrangement with its own terms.
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 · Self-hosted · Linux
- Website
- learn.microsoft.com
- Requires an account
- Yes
- Free tier
- Yes
- Watermark on the free tier
- Not verified
- Maximum file size
- 500 MB
- Batch processing
- Yes
- Data residency
- Multiple regions
- Open source
- No
- Self-hostable
- Yes
- Licence
- Not verified
- Redaction
- No redaction feature
- Programmatic access
- HTTP API
- Last availability check
- Responded in 142 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.
https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/overview
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.
-
“Azure containers aren't licensed to run without being connected to Azure for metering.”
Checked by us - the source says this Read 22 September 2026
Confirmed verbatim, and the sentence that follows it is the one that matters to a reader weighing a container deployment: "Customers need to enable the containers to always communicate billing information with the metering service. Azure containers don't send customer data (for example, the image or text that is being analyzed) to Microsoft." So the required connection is for metering rather than for the document, which is a materially better position than the claim alone suggests - and the same page makes plain that it is not optional.
The practical consequence is that an air-gapped deployment is not permitted under this licence, which is the question anyone considering containers for sensitive documents is actually asking. Recorded as the vendor's description; we have not run a container 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
- Read
- https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/containers/install-run
- From
- the text a browser renders
- SHA-256
- 5c8ad17f6d5144cf1a6e2c0a3d24a04f397bf006eff915e4d891616c6c7fa1ca
How we checked
Re-verified verbatim on 2026-09-22: https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/containers/install-run 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-learn-microsoft-com-en-us-azure-ai-services-document-intel-be74f159.txt.
-
“The incoming data is processed in the same region where the Document Intelligence resource was created.”
Checked by us - the source says this Read 22 September 2026
Confirmed verbatim. This is a data-residency claim and it is the most useful kind of statement in this tier because it says where processing happens rather than how good the product is.
Read with the neighbouring claim about container metering, verified earlier, the two together give a coherent picture: the container must report billing to Azure, documents are processed in the resource's own region, and per that earlier page the images and text being analysed are not sent to Microsoft. The page does not state whether any metadata leaves the region, and that was not pursued.
Recorded as the vendor's documentation; nothing was deployed and no control was run.
Standing: Re-read and matched once per-response tokens were masked
Check it yourself: the source, the SHA-256, and how we checked
- From
- the text a browser renders
- SHA-256
- 41ac82e782756e446dabd1392604859c3a3ce448bc1f2abee8075e382b3a83c9
How we checked
Re-verified verbatim on 2026-09-22: https://learn.microsoft.com/en-us/azure/foundry/responsible-ai/document-intelligence/data-privacy-security 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-learn-microsoft-com-en-us-azure-foundry-responsible-ai-doc-9ea06fd7.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.