Known limitations to Document AI¶
This topic describes known limitations to Document AI.
Document AI has the following limitations:
Document AI supports processing documents in English. The results for other languages might not be satisfactory.
Document AI supports processing documents of specific formats and sizes.
For more information, see Prepare your documents for Document AI.
Document AI supports processing a maximum of 1000 documents in one query.
Document AI does not support a whole table extraction in a single query.
Document AI does not support multiple users working on the same model build at the same time in Snowsight. For example, two users cannot upload documents or review answers within the same model build at the same time.
Document AI does not support double quotes around identifiers for the database and schema.
Document AI does not support altering a database or a schema where the model build is located.
Document AI does not support serverless tasks.
Document AI is available to accounts in AWS and Microsoft Azure commercial regions, with some exceptions. For more information, see Document AI availability.
Document AI availability¶
Document AI is currently available in the following regions:
Cloud platform |
Cloud region |
---|---|
Amazon Web Services (AWS) |
|
Microsoft Azure |
|
For more information, see Supported Cloud Regions.
If you want to process documents stored in a cloud platform or region that is not supported by Document AI, create an account in a supported Snowflake region, and then create an external stage to connect to the cloud storage provider that contains your documents. For more information, see Overview of data loading.