Sic Semper Errata / Docs
Browse docs · File types & OCR

SSE user guide

File types and OCR

Check supported uploads, scan processing, and the roadmap for Office documents, images, spreadsheets, and email.

Start with PDF or plain text in the public demo. Word documents, spreadsheets, email files, and direct image uploads are on the roadmap.

Available in the public demo

Available now

PDF .pdf

Import an incoming request, or collect a responsive record for visual review and region redaction. Keep the original and inspect the separate release copy.

Available now

Plain text .txt

Collect a responsive record encoded as UTF-8. Review the text and redact exact, case-sensitive phrases. TXT is not an incoming-request import format.

Public limits: 1 MB per file. PDFs may contain up to 3 pages, each no larger than 17 × 17 inches. Encrypted or unreadable PDFs are rejected.

Incoming request PDFs need embedded text. A scanned PDF can be collected as a responsive record and reviewed visually, but that workflow does not extract text with OCR.

PDF release copies are rebuilt from page images; the original searchable text is not retained. TXT release copies remain text. See review and release for the approval steps.

OCR: reading text from a scan

OCR is a processing feature, not a file extension. It turns the words visible in a scan into extracted text that staff can check against the original.

Direct TIFF, BMP, PNG, and JPEG uploads are planned. Current OCR does not imply support for those uploads, handwriting recognition, other languages, or automatic exemption decisions.

Local application file limits

Incoming request PDFs: 15 MB and 25 pages. Responsive PDF and TXT records: 10 MB; responsive PDFs are limited to 20 pages and 17 × 17 inches per page. These are local application limits, not the public demo limits. Local-folder search supports PDF and TXT; live vendor connectors remain planned.

File-format roadmap

These formats are planned, not accepted by the current upload controls. The order below is the proposed implementation sequence; no release dates are committed.

  1. Planned · scans and images

    TIFF, BMP, PNG & JPEG

    .tif, .tiff, .bmp, .png, .jpg, .jpeg

    Import single images and multipage TIFFs, preview every page, and extend OCR to collected scans. Keep extracted text separate from the original and require review of OCR errors.

  2. Planned · documents

    Word, rich text & OpenDocument

    .doc, .docx, .rtf, .odt

    Retain the native document and create a review copy. Make comments, tracked changes, and embedded content visible for review before approving a release.

  3. Planned · spreadsheets

    Excel, CSV & OpenDocument sheets

    .xls, .xlsx, .csv, .ods

    Review sheets and cells, including hidden content, comments, and formulas. Approve what is released in the native file or a separate review copy; do not silently treat a workbook as a PDF.

  4. Planned · email

    Email messages & attachments

    .eml, .msg

    Preserve message headers, body, timestamps, and attachment relationships. Review each attachment under its own file-type rules. File import is separate from a live Microsoft 365 or Google Workspace connection.

  5. Later · presentations and bundles

    PowerPoint, OpenDocument slides & ZIP

    .ppt, .pptx, .odp, .zip

    Inspect slides, notes, and embedded files. For archive imports, list and review each member with extraction limits. The demo already exports an approved ZIP package; it does not import ZIP archives.

What counts as a supported format?

Accepting an upload is only the first step. Each format must preserve the original, show a usable review copy, support an explicit release decision, and produce a verified output without leaking removed content. Unsupported, damaged, or password-protected files need a clear exception path. Planned OCR expansion includes review of extraction errors; handwriting and additional languages need separate validation.

Can I rename an unsupported file to PDF or TXT?

No. Changing the extension does not convert the file. For a fictional demo, export a supported PDF or plain-text copy in the source application and inspect it before uploading. Keep the original separately; conversion can omit comments, attachments, or other content.