Home / AI Service Directory / Google Document AI

Google Document AI

Google · Data analysis · US · cloud.google.com

Last verified 2026-09-16
Entry services/google-document-ai.yaml
Training on your data (default tier)
Customer documents and custom models are private customer data and are never used to train or improve Document AI models
Retention
Online requests processed in memory and not persisted; batch documents deleted after processing with a one-day failsafe
Data residency
Regional and multi-regional endpoints; the customer selects the processing location

Vendor data-handling terms

Training on your data (default tier)
Customer documents and custom models are private customer data and are never used to train or improve Document AI models
At Google Cloud, we never use customer data to train our Document AI models. / Documents and custom models are considered to be (private) customer data.
checked 2026-09-16
Google Cloud Document AI, Security and data governance
Retention
Online requests processed in memory and not persisted; batch documents deleted after processing with a one-day failsafe
For online (immediate response) operations, the document data (sent in the request) is processed in memory, encrypted in flight, and not persisted to disk. / For batch operations, the stored document is typically deleted immediately after the processing, with a failsafe Time to live (TTL) of one day.
checked 2026-09-16
Google Cloud Document AI, Security and data governance
Data residency
Regional and multi-regional endpoints; the customer selects the processing location

EU and US multi-regions are documented in the Document AI regions page.

Regional and multi-regional support
checked 2026-09-16
Google Cloud Document AI, Security and data governance

Domains and endpoints

observed means seen in Unseen deployments; vendor-documented means listed by the vendor. Vendors do not publish complete lists.

HostRoleSource
cloud.google.comappobserved
documentai.googleapis.comappobserved

Assessment

API: low

Reasoning: Google Cloud's document parsing and OCR API. No training on customer documents, in-memory processing for online calls, one-day ceiling for batch, and a customer-chosen region, all under the Google Cloud data processing terms. As good as a hosted document API gets.

Assessed 2026-09-16. The assessment is Unseen's; the terms above are the vendor's.

Changelog

  • 2026-09-16Documented. Training, retention and region terms verified against the Document AI security page.
  • 2026-09-15Listed from the Unseen catalogue with observed domains.

Corrections

Pull request on GitHub, or the form below. Changes are reviewed and recorded in the changelog.

Edit on GitHub