Skip to main content
ABBYY Vantage has the following limitations and known issues. ABBYY plans to address them in future releases.

​
Files, pages, and browsers

  • A transaction can contain up to 1,000 files.
  • If you upload several files through the Public API into a transaction with an OCR, Document, or Classification skill, Vantage merges and processes them as one document.
  • Each imported file can be up to 2 GB.
  • Unattended document processing has been tested with up to 3,000 pages for an OCR skill and up to 500 pages for a Document skill. For Document and Classification skills in a processing flow with a manual review stage, it has been tested with up to 100 pages. Documents with more pages can slow down your work in the Manual Review client.
  • Vantage has been tested only with Google Chrome and Safari. Firefox, Opera, and other browsers aren’t fully supported yet, and you may run into issues with them. For more information, see Supported browsers.
  • Automatic skill selection for files that you upload to Documents works only for English, French, German, and Spanish.
  • You can import Excel files (*.xls, *.xlsx) with the following limitations:
    • If a file contains a chart with text, such as a title, Vantage doesn’t extract that text.
    • Vantage picks a page size from a set of standard sizes. If the contents of a file don’t match any standard size (for example, they’re too wide), the file may display incorrectly.
    • If a file contains slicers with data outside the visible area (for example, items that you have to scroll to), Vantage doesn’t display that data.
    • Vantage ignores the Excel Scale to Fit option, which shrinks a worksheet to fit the printed page. If a table has too many rows, Vantage splits it across multiple pages.
    • Vantage may not keep the position of a table relative to an object outside the table. For example, text in the table can shift to the right relative to a shape above it.

​
Process skill

  • If a Process skill contains more than one Classify or Extract activity, Online Learning applies only to the last of them.
  • You can’t place Manual Review, Input, Output, Assemble, or For Each Document activities inside a For Each Document activity.
  • Each ZIP archive that Vantage imports from a shared folder can be up to 1.5 GB, and each file inside it can be up to 100 MB. The archive must contain only files in supported formats. Otherwise, the import fails.
  • The Custom activity supports only synchronous Vantage API calls. In a synchronous call, the external system processes the request in a few seconds and returns the results in its response almost instantly.
  • One Extract activity supports up to 50 Document skills.
  • When Vantage imports from email, it processes only first-level attachments as transaction documents. If an email has another email attached, Vantage doesn’t import the attachments of that attached email.
  • When Vantage imports from email, it sets the FileName document registration parameter but not the FilePath parameter.
  • You can’t export image fields to CSV. If a Document skill has an image field, its value is always empty in the CSV export results.
  • When you specify skill parameters with the API, a transaction can have up to 10 key-value pairs. Each key can be up to 128 characters, and each value up to 256 characters.

​
OCR skill and OCR activity

  • Selecting more than 20 recognition languages can slow down performance and impair the OCR skill.
  • The maximum recommended number of pages per file for an OCR skill is 3,000.
  • Objects such as stamps and stickers can sometimes appear skewed in output PDFs.

​
Document and Classification skills

  • Training sets can contain up to 10,000 documents.
  • When you add new documents to a trained Document or Classification skill, the Results tab shows outdated information until you label the new documents and retrain the skill.
  • When you compare training results to the reference in a Document skill, the Difference tab doesn’t show the error type and reference value for repeating fields.
  • A skill can have up to 100 parameters.

​
Document Splitter skill

  • The Document Splitter skill doesn’t support reference labeling and statistics yet.
  • You can’t revert the Document Splitter skill to a previous version.
  • Documents that you process with the Document Splitter skill shouldn’t contain more than 1,000 pages each.

​
Online Learning

  • Online Learning doesn’t support the following types of fields:
    • Image fields
    • Repeating groups that contain image fields
    • Nested repeating groups
    • Groups inside repeating groups or tables
  • Online Learning results for checkmark fields are generally high quality. Sometimes, though, checkmarks aren’t detected or trained properly. This can happen because of the checkmark field’s shape (for example, round or square). It can also happen because of how the user marked the field (for example, crossing it out or underlining it instead of placing a checkmark inside).
  • The training set can contain up to 10,000 documents, and the testing set up to 1,000.

​
Fast Learning in Advanced Designer

  • Deleting a Fast Learning activity from a skill also deletes the database lookup rules.

​
Deep Learning in Advanced Designer

  • Deep Learning works only with text fields.
  • One Deep Learning activity can be trained for up to 50 fields, including repeating groups and fields, or for one table of up to 32 columns.
  • You can’t train nested repeating groups or groups inside repeating groups.

​
Data catalogs

  • You can’t delete or change a data catalog that a skill uses, in any version of that skill. To delete a data catalog that you no longer need, stop using it in the skill by deleting or changing the corresponding rule. Then delete all previous versions of the skill. For more information, see Delete skill versions. To change the data catalog but keep the old skill versions, create a new data catalog.
  • A field value isn’t filled in automatically when a lookup rule is used for one data catalog. To fill in the value, click the Lookup icon for the field, select the value from the lookup results, and then retrain the skill.

​
Skill Monitor

  • You can export only the latest 10,000 errors from the Error Log to a CSV file. To reduce the number of exported errors, filter the Error Log or search for specific lines.
  • If you delete a skill and then import it again with the same identifier, Skill Monitor shows no metrics for it.

​
Vantage API

  • When you get a list of transactions, the request fails if the Offset parameter equals totalItemCount. For more information, see Getting a list of transactions.

​
Advanced Designer

  • The Number must be within interval property is disabled for Number and Money fields in the Manage Fields dialog box. You can set this property in the Reference section of the Fields tab or in Vantage.
  • Field settings that you edit and save in the Manage Fields dialog box aren’t synchronized with the reference labeling section. To work around this, edit field settings in Vantage or in the Reference section of the Fields tab in Advanced Designer.

​
Built-in skills

  • Built-in skills process documents only in their supported languages. In a copy of a built-in skill, enabling additional languages doesn’t let the skill extract fields from documents in those languages. You can disable any of the supported languages. Disable the ones that your documents don’t use.
  • For the Identity Document skill, Manual Review doesn’t show the regions of extracted fields.

​
Mobile Input

  • Passport capture isn’t automated. Center the document on the screen, and then capture the image manually.

​
Scanning Station

If you install Scanning Station without administrator permissions, it installs with the English interface, even if you selected a different interface language during setup. To get the interface language that you selected, do one of the following:
  • Get administrator permissions, and then install Scanning Station.
  • Reinstall Scanning Station.
  • After you install Scanning Station, change the following registry key. Replace X with the code for your language from the following table.