This month's release includes two features in early access, accuracy improvements for smart learning, and a handful of fixes.

Bounding Box in Extraction Results (Early Access)

Detection region fields now return a `bounding_box` alongside the extracted value. The bounding box tells you where on the document FormX found each piece of data: the coordinates of the region that was captured.

A few things this makes easier:

• If you're building a review UI where a human confirms extracted values, you can highlight the exact region on the source document instead of asking reviewers to hunt for the value manually.
• When an extraction looks wrong, checking the bounding box tells you what region the model was looking at.
• If you're passing extracted data to another system, you can carry positional context along with the value.

This feature is currently available to selected users. Broader rollout is planned. Reach out if you'd like early access.

Azure Claude Model Support (Enterprise)

FormX now supports Claude models hosted on Azure through AnthropicFoundry. If your infrastructure runs on Azure and you've been waiting to use Claude in production, this removes the main blocker.

Previously, using Claude with FormX required a separate Anthropic API setup. With AnthropicFoundry support, teams on Azure can connect directly without managing an additional integration point.

Available on Enterprise plans. Contact us to learn more.

Smart Extraction for Smart Learning

Smart learning now applies training samples in a consistent order when building prompts. Extraction results are more stable and predictable across similar documents.

If you've been using smart learning and noticed variability in results, this change should help.

Other Changes

• Improved passport MRZ field extraction
• Troubleshooting improvements for teams running parallel UAT and production environments
• Misc bug fixes and security updates

Questions? Reach out through the FormX portal or contact your account manager.