Google Cloud

The Document AI logo. The left side is a document icon, the right is circuitry spreading out.

Document AI, parse, and process documents at scale

Document AI lets developers create high-accuracy processors to extract unstructured or structured data from documents, classify, and split documents, automating tedious tasks.

Features

Custom extractor

Custom extractor provides an easy way to extract structured data from documents. Custom extractor is powered by generative AI, which means it can be used out of the box to get accurate results across a wide array of documents. Furthermore, you can achieve higher accuracy by providing as few as 10 documents to fine-tune the large model—all with a simple click of a button or an API call.

Custom splitter

Custom splitter is designed to split composite documents (documents made up of multiple classes) into a number of single class documents by identifying each logical document. For example, a mortgage package contains multiple classes within it such as application, income verification, and photo ID. Custom splitter processors can be used out of the box, or trained from the ground up using your own documents and custom classes.

Custom classifier

Use custom classifier to classify documents. Build it from the ground up with your own documents and custom classes. Its generative AI aspect allows few-shot learning and fine-tuning. These improve accuracy with fewer samples and corrections with iterative auto-labeling.

OCR parser

You can use Enterprise Document OCR as part of Document AI to detect and extract text and layout information from various documents. With configurable features, you can tailor the system to meet specific document-processing requirements.

How It Works

Use the Google Cloud Console to select the parser that is right for your project needs. With workflows enabled, you can automate document ingestion, processing, and storage in a seamless process.

Common Uses

Find insights in documents with BigQuery

Tutorials, quickstarts, & labs

Integrate with BigQuery to extract and use metadata

You can now extract metadata from documents directly into a BigQuery objects table. Seamlessly join the parsed data with other BigQuery tables to combine structured and unstructured data, paving the way for comprehensive document analytics.

Integrate with BigQuery to extract and use metadata

You can now extract metadata from documents directly into a BigQuery objects table. Seamlessly join the parsed data with other BigQuery tables to combine structured and unstructured data, paving the way for comprehensive document analytics.

Digitize text for ML model training

Tutorials, quickstarts, & labs

Extract value from archives with Enterprise Document OCR

Enterprise Document OCR enables users to create value from archival content that is otherwise unusable for AI and ML training. OCR extracts text from scanned documents, plots, reports, and presentations prior to saving on a cloud storage or a data warehouse. Use these high-quality OCR outputs to boost your digital transformation initiatives such as training ML models specific to your business.

Extract value from archives with Enterprise Document OCR

Enterprise Document OCR enables users to create value from archival content that is otherwise unusable for AI and ML training. OCR extracts text from scanned documents, plots, reports, and presentations prior to saving on a cloud storage or a data warehouse. Use these high-quality OCR outputs to boost your digital transformation initiatives such as training ML models specific to your business.

Generate a solution

What problem are you trying to solve?

What you'll get:

Step-by-step guide

Reference architecture

Available pre-built solutions

Pricing

How Document AI pricing worksCost is based on number of processed pages per month, the relevant quota, and any purchased capacity reservation.
CategoryParserPrice

Digitize text

Enterprise Document OCR processor

$1.50 per 1,000 pages

1 - 5,000,000 pages per month

OCR add ons

$6 per 1,000 pages

1 - 5,000,000 pages per month

Extract structures and entities from documents

Custom extractor

$30 per 1,000 pages

1 - 1,000,000 pages per month

Form parser

$30 per 1,000 pages

1 - 1,000,000 pages per month

Layout Parser (Includes initial chunking)

$10 per 1,000 pages

1 - 1,000,000 pages per month

Classify documents

Custom splitter

$5 per 1,000 pages

1 - 1,000,000 pages per month

Custom classifier

$5 per 1,000 pages

1 - 1,000,000 pages per month

Summarizer

$25 per 1,000 pages

1 - 1,000,000 pages per month

How Document AI pricing works

Cost is based on number of processed pages per month, the relevant quota, and any purchased capacity reservation.

Digitize text

Parser

Enterprise Document OCR processor

Price

$1.50 per 1,000 pages

1 - 5,000,000 pages per month

OCR add ons

Parser

$6 per 1,000 pages

1 - 5,000,000 pages per month

Extract structures and entities from documents

Parser

Custom extractor

Price

$30 per 1,000 pages

1 - 1,000,000 pages per month

Form parser

Parser

$30 per 1,000 pages

1 - 1,000,000 pages per month

Layout Parser (Includes initial chunking)

Parser

$10 per 1,000 pages

1 - 1,000,000 pages per month

Classify documents

Parser

Custom splitter

Price

$5 per 1,000 pages

1 - 1,000,000 pages per month

Custom classifier

Parser

$5 per 1,000 pages

1 - 1,000,000 pages per month

Summarizer

Parser

$25 per 1,000 pages

1 - 1,000,000 pages per month

Pricing calculator

Estimate your monthly costs, including region specific pricing and fees.

Custom quote

Connect with our sales team to get a custom quote for your organization.

Start your proof of concept

Get started with a $300 credit

Want to learn more about Document AI?

Process a document

Train a custom processor

Evaluate processor performance

Partners & Integration

Document AI partners

Other inquiries and support

FAQ

How do we differentiate?

Generative AI - we provide simple access to powerful foundation models that help our customers create parsers to extract documents with a simple journey and in minutes. This solves two pain points: users do not need to label or prepare datasets for their custom models, and users do not need to worry about specifics like converting document types, choosing models, few shot samples, or chunking.

In what regions is Document AI available?

How do I monitor performance?

Precision, recall, F1 score, and more for each parser can be monitored directly from the Google Cloud Console, with a specific user interface to visualize loads and performance. You can learn more in the evaluate page.

Can I increase pages processed or reserve processing capacity?

Yes, you can increase the quota for a project, increasing the number of pages processed per minute.

You can also make capacity reservation requests for periods of high volume traffic.

Read the original on cloud.google.com ↗