Try Gemini 1.5 models, the latest multimodal models in Vertex AI, and see what you can build with up to a 2M token context window. Try Gemini 1.5 models, the latest multimodal models in Vertex AI, and see what you can build with up to a 2M token context window.

Pretrained overview

Document AI offers multiple products to process documents for information for different use cases.

Pretrained parsers

For more information, go to Explore pretrained processors.

Bank statement parser

Bank statement parser extracts key-value pairs (KVP). It can extract up to 17 generic entities. Examples include: Account number, client name, bank name, and table items like deposits and withdrawals. You don't specify the fields (schema) you want to extract. Bank statement parser supports Enrichment and Normalization.

W2 parser

W2 parser extracts from the IRS Form W2 as KVP. It can extract up to 12 generic entities, including employee name, Social Security Number, employer, and wages. You don't specify the fields (schema) you want to extract. W2 parser supports Enrichment.

US passport parser

US passport parser extracts KVP. It can extract up to seven generic entities. These include given names, family names, document ID, and date of birth. You don't specify the fields (schema) you want to extract. US passport parser supports Normalization.

Utility parser

Utility parser extracts KVP. It can extract up to 75 generic entities from utility bills. These include supplier name, previous paid amount, and line items like amount, description, and product code and quantity. You don't specify the fields (schema) you want to extract with the utility parser.

Identity document proofing parser

Identity document proofing parser predicts the validity of ID documents using multiple signals.

fraud_signals_is_identity_document detection: Predicts whether an image contains a recognized identity document.
fraud_signals_suspicious_words detection: Predicts whether words are present that aren't typical on IDs.
fraud_signals_image_manipulation detection: Predicts whether the image was altered or tampered with an image editing tool.
fraud_signals_online_duplicate detection: Predicts whether the image can be found online (US only).

Pay slip parser

Pay slip parser extracts KVP. It can extract up to 26 generic entities from pay slips. These include employee name, bonus, commissions, overtime, and pay date. You don't specify the fields (schema) you want to extract. Pay slip parser supports Enrichment and Normalization.

US driver license parser

US driver license parser extracts KVP. It can extract up to eight generic entities from a driver license. Examples include: Given name, family name, document ID, and expiration date. You don't specify the fields (schema) you want to extract. US driver license parser supports Normalization.

Expense parser

Expense parser extracts KVP. It can extract up to 17 generic entities from expense reports. Examples include: Expense date, supplier name, total amount, and currency. You don't specify the fields (schema) you want to extract. Expense parser supports Enrichment and Normalization.

Invoice Parser

Invoice Parser extracts KVP. It can extract up to 46 generic entities from invoices. These include invoice number, supplier name, invoice amount, tax amount, invoice date, and due date. You don't specify the fields (schema) you want to extract. Invoice Parser supports Enrichment and Normalization.

Summarizer

Summarizer gives abstract and bullet point summaries for short and long documents. Summarizer also lets you specify the output length of the summary as comprehensive, medium, or brief.

Layout Parser