Mantenha tudo organizado com as coleções
Salve e categorize o conteúdo com base nas suas preferências.
A API v1beta2 da Document AI foi descontinuada e substituída pela API
v1.
Ao contrário de v1beta2, a v1 da Document AI é uma API com estado. Isso significa que você precisa criar um Processor antes de enviar solicitações aos documentos Process e BatchProcess.
OBSERVAÇÃO: essa descontinuação afeta apenas a versão da API usada para chamar qualquer processador.
Isso não afeta a disponibilidade de nenhum processador.
As principais áreas de incompatibilidade da API v1 com a API v1beta2 são discutidas mais adiante neste documento.
Como criar um processador
Para migrar para v1 na criação de processadores, primeiro crie um Processor. A tabela a seguir serve como uma orientação geral para escolher um ProcessorType adequado para criar, mas você pode consultar a lista completa de processadores.
Para ver um guia explicativo sobre a criação de um processador, consulte o
guia.
Processando documentos
A Document AI v1 tem duas APIs para aceitar documentos: Process e BatchProcess.
Cada um desses métodos corresponde a um método de v1beta2. Os métodos e as diferenças entre eles são destacados:
Além de aceitar documentos individuais no Cloud Storage, o v1 é compatível com um prefixo do Cloud Storage para processar vários documentos de uma só vez.
Bibliotecas de cliente
Se você estiver usando o Client Libraries para interagir com a
Document AI, mude do uso do DocumentUnderstandingService para o
DocumentProcessorService. Para um exemplo detalhado de como usar as bibliotecas de cliente
para v1, consulte bibliotecas de cliente.
[[["Fácil de entender","easyToUnderstand","thumb-up"],["Meu problema foi resolvido","solvedMyProblem","thumb-up"],["Outro","otherUp","thumb-up"]],[["Difícil de entender","hardToUnderstand","thumb-down"],["Informações incorretas ou exemplo de código","incorrectInformationOrSampleCode","thumb-down"],["Não contém as informações/amostras de que eu preciso","missingTheInformationSamplesINeed","thumb-down"],["Problema na tradução","translationIssue","thumb-down"],["Outro","otherDown","thumb-down"]],["Última atualização 2025-09-04 UTC."],[[["\u003cp\u003eThe Document AI \u003ccode\u003ev1beta2\u003c/code\u003e API is deprecated and users should migrate to the \u003ccode\u003ev1\u003c/code\u003e API.\u003c/p\u003e\n"],["\u003cp\u003eUnlike \u003ccode\u003ev1beta2\u003c/code\u003e, the \u003ccode\u003ev1\u003c/code\u003e API is stateful and requires the creation of a \u003ccode\u003eProcessor\u003c/code\u003e before processing documents.\u003c/p\u003e\n"],["\u003cp\u003eWhen migrating to \u003ccode\u003ev1\u003c/code\u003e, the method to process documents is either \u003ccode\u003eProcess\u003c/code\u003e or \u003ccode\u003eBatchProcess\u003c/code\u003e which each corresponds to a method from \u003ccode\u003ev1beta2\u003c/code\u003e respectively, but with some slight differences.\u003c/p\u003e\n"],["\u003cp\u003eTo migrate to \u003ccode\u003ev1\u003c/code\u003e for processor creation, you will first need to create a \u003ccode\u003eProcessor\u003c/code\u003e and use a \u003ccode\u003eProcessorType\u003c/code\u003e based on the document type.\u003c/p\u003e\n"],["\u003cp\u003eUsers of the Document AI client libraries must switch from \u003ccode\u003eDocumentUnderstandingService\u003c/code\u003e to \u003ccode\u003eDocumentProcessorService\u003c/code\u003e to use the \u003ccode\u003ev1\u003c/code\u003e API.\u003c/p\u003e\n"]]],[],null,["Document AI's `v1beta2` API is now deprecated in favor of the\n[`v1`](/document-ai/docs/reference/rest/v1/projects.locations.processors) API.\n\nUnlike `v1beta2`, Document AI [`v1`](/document-ai/docs/reference/rest/v1/projects.locations.processors)\nis a stateful API, meaning you need to create a [`Processor`](/document-ai/docs/reference/rest/v1/projects.locations.processors#resource:-processor)\nbefore you can send requests to [`Process`](/document-ai/docs/reference/rest/v1/projects.locations.processors/process)\nand [`BatchProcess`](/document-ai/docs/reference/rest/v1/projects.locations.processors/batchProcess)\ndocuments.\n\nNOTE: This deprecation only affects the version of the API used to call any processor.\nIt does not affect the availability of any processor.\n\nThe major areas of v1 API backward incompatibility with v1beta2 API are discussed\nlater in this document.\n\nCreating a processor\n\nTo migrate to `v1` for processor creation, first create a [`Processor`](/document-ai/docs/reference/rest/v1/projects.locations.processors#resource:-processor). The following table serves as a general guideline for choosing an appropriate\n[`ProcessorType`](/document-ai/docs/reference/rest/v1/projects.locations.processorTypes#resource:-processortype)\nto create, but you can consult the full [processor list](/document-ai/docs/processors-list).\n\n| `document_type` | `params` | Processor type |\n|-----------------|----------------------------|--------------------------------------------------------------------------------------|\n| `invoice` | | [`INVOICE_PROCESSOR`](/document-ai/docs/processors-list#processor_invoice-processor) |\n| `general` | | [`OCR_PROCESSOR`](/document-ai/docs/processors-list#processor_doc-ocr) |\n| `general` | `table_extraction_params` | [`FORM_PARSER_PROCESSOR`](/document-ai/docs/processors-list#processor_form-parser) |\n| `general` | `form_extraction_params` | [`FORM_PARSER_PROCESSOR`](/document-ai/docs/processors-list#processor_form-parser) |\n| `general` | `entity_extraction_params` | [`FORM_PARSER_PROCESSOR`](/document-ai/docs/processors-list#processor_form-parser) |\n| `general` | `automl_params` | [`CUSTOM_EXTRACTION_PROCESSOR`](/document-ai/docs/processors-list#processor_cde) |\n\nFor a step-by-step guide to creating a processor, view the\n[guide](/document-ai/docs/create-processor).\n\nProcessing documents\n\nDocument AI `v1` has two APIs to accept documents: [`Process`](/document-ai/docs/reference/rest/v1/projects.locations.processors/process) and [`BatchProcess`](/document-ai/docs/reference/rest/v1/projects.locations.processors/batchProcess).\nEach of these methods corresponds to a method from `v1beta2`. The methods and\ntheir differences are highlighted:\n\n| `v1beta2` Method | `v1` Method | Migration notes |\n|-------------------------|--------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------|\n| `ProcessDocument` | [`Process`](/document-ai/docs/reference/rest/v1/projects.locations.processors/process) | `InputConfig.Cloud Storage_source` not supported in `v1`, use `BatchProcess` to process documents on Cloud Storage. |\n| `BatchProcessDocuments` | [`BatchProcess`](/document-ai/docs/reference/rest/v1/projects.locations.processors/batchProcess) | In addition to accepting individual documents on Cloud Storage, `v1` supports a Cloud Storage prefix to process many documents at once. |\n\nClient libraries\n\nIf using the [`Client Libraries`](/document-ai/docs/libraries) to interact with\nDocument AI, switch from using the `DocumentUnderstandingService` to the\n`DocumentProcessorService`. For a detailed example of using the client libraries\nfor `v1`, see [client libraries](/document-ai/docs/libraries)."]]
