This release focuses on three things: making it easier to navigate mixed document batches in Workspace, improving the reliability of AI extraction, and reducing errors when processing passport MRZ fields.
Documents grouped by type in Workspace
A common scenario: you receive a single PDF with dozens of documents inside — some invoices, some delivery notes, some identity documents — all mixed together. FormX's Split and Classify feature handles this by automatically splitting the file and identifying each document type. This release extends that further in the Workspace UI.
Results are now grouped by document type. A document type column shows the classification for each file, and you can filter the list by type and processing status. If you want to review all the invoices in a batch, or find the ones flagged for review, you can filter to that view directly instead of scanning a flat list.
For teams processing high volumes of mixed documents, this makes triage significantly faster.
More reliable AI extraction
FormX has added newer AI models and improved how the extraction engine handles edge cases. One common failure mode is when a model call succeeds at the API level but returns an empty or incomplete response — these would previously fail silently. FormX now retries in these cases and surfaces clearer signals when results may be unreliable.
The default invoice extractor has also been updated to a more accurate model, improving extraction quality on complex or varied invoice layouts.
Better passport MRZ parsing
Machine-readable zone (MRZ) extraction from passports has been improved to handle a wider range of real-world documents. The parser now does a better job with ambiguous OCR characters and edge cases in certain passport formats.
If you process international identity documents and have seen occasional MRZ extraction errors, this should reduce them.
Other changes
- Bug fixes
