File types and OCR

Use PDF for incoming requests and PDF or UTF-8 text for responsive records. The public demo stays on those formats. More formats have extraction components in development; their complete review and release workflows are not verified, and they are not offered on the public demo.

1. Choose the workflow below: incoming request or responsive record.
2. Check the public-demo or reference-application limits for that workflow.
3. Inspect the original and the release copy before approving a package.

Supported workflows

The reference application runs locally with the required processing tools. Its limits are different from the public demo. Verified upload, review, and release remain PDF and UTF-8 TXT.

Public PDFs and local responsive PDFs have a 17 × 17-inch page limit. Encrypted or unreadable PDFs and invalid UTF-8 text are rejected. TXT is not a verified incoming-request format. The public demo does not accept presentations, email, Office documents, images, or ZIP uploads.

A scanned PDF can be collected as a responsive record for visual review. The application’s record-review interface does not run OCR. PDF release copies are rebuilt from page images and do not retain the original searchable text. TXT release copies remain text.

Catalog-supported extract

The reference file catalog marks PowerPoint .pptx as supported for text extraction: slide text, speaker notes, and hidden-slide flags. That is extract-only. It is not a reviewed redaction or release workflow, and it is not available in the public demo.

Email .eml extraction and ZIP inventory modules exist in the reference codebase. Both remain planned in the catalog. Do not treat a catalog row, a file picker, or an extractor as a complete workflow.

OCR: text from a scan

OCR is a processing step, not a file extension. It reads words visible in a scan so staff can compare extracted text with the original.

Image, Office, MSG, and connector workflows are not available. An extractor or API route is not a public-demo format.

Planned formats

These workflows are not verified for use in the public demo or as complete reference-application review and release, even where a picker lists their extensions or an extractor exists. No release dates are committed.

1. Planned · scans and images

   TIFF, BMP, PNG & JPEG

   .tif · .tiff · .bmp · .png · .jpg · .jpeg

   Connect direct image and multipage TIFF import, page previews, and the OCR API to a reviewed interface.

2. Planned · documents

   Word, rich text & OpenDocument

   .doc · .docx · .rtf · .odt

   Keep native originals and make comments, tracked changes, and embedded content visible before release. Office formats are not available.

3. Planned · spreadsheets

   Excel, CSV & OpenDocument sheets

   .xls · .xlsx · .csv · .ods

   Review sheets, cells, formulas, and hidden content. Approve the chosen native or rendered output.

4. Planned · email

   Email messages & attachments

   .eml · .msg

   An EML extract module exists; the catalog still lists EML as planned. MSG is not available. File import is separate from a live mail connector.

5. Catalog extract · presentations

   PowerPoint text extract

   .pptx

   Catalog-supported text extract only. PPT and ODP remain planned. Slide review, media rendering, redaction, and release are not verified. Not offered on the public demo.

6. Planned · other slides and bundles

   Legacy slides, OpenDocument slides & ZIP

   .ppt · .odp · .zip

   A ZIP inventory module exists; the catalog still lists ZIP as planned. Approved ZIP export already exists and does not establish archive import.

Follow review and approval steps or check planned system connections.