VIEW-008 · EDIT-001 · EDIT-002 · OCR-001
Selecting, adding and editing text
Choose Select text (or Pages) to copy searchable source text locally; copy restrictions are honored. Text → Edit existing text supports bounded printable ASCII runs in unambiguous original fonts, retaining font, size and placement without overlays or reflow. Replacements must fit and require actual-PDF review. Text → OCR this page runs English/French recognition locally on unrotated, image-only pages; it adds invisible search/copy text without changing the scan visually. OCR is not automatic, can misread words, and does not make scanned words visually editable. Unknown glyphs, ambiguous accessibility text and unsafe structures refuse editing. Added overlays share point sizes, wrapping, alignment and baselines with export; system glyph outlines can vary. No universal PDF editing, OCR accuracy or accessibility certification is claimed.
SEC-001 · SEC-002 · SEC-003
Protection copies
Add-password and authorized unlock create new local copies within 50 MiB and 500 pages. New passwords support 12–127 printable ASCII characters without substitution. Signed, active, attached-file, XFA, damaged/incremental and unsupported string layouts stop without an export; adding protection also refuses PDF/A and existing extensions. Content, metadata and source objects are verified before returning a copy. Passwords clear after each attempt; authority resets for each source, and changing settings discards results. JavaScript cannot guarantee memory erasure. Inspect saved copies in your reader. Development-gated permission controls use separate owner/open passwords and explicit setting review. An empty open password lets anyone open and extract the file. Flags are reader-enforced requests, not unbreakable DRM, and accessibility extraction is always allowed without certifying accessibility. Production/distribution defaults remain off. No broad-reader or native release certification is claimed.
CONV-005
Plain-text conversion
One UTF-8 .txt file up to 1 MiB supports printable ASCII, tabs and line breaks only. Invalid encoding, unsupported characters or layouts over 500 pages stop without substitution or truncation. A4/Letter portrait uses Courier 12 pt, 36 pt margins and 15 pt line spacing. Tabs expand at four-column source-line stops, long lines wrap and blank lines remain; markup is literal. Review the actual output before saving. The selectable-text PDF is untagged, not an accessible PDF certification or byte-identical text archive. Rich text, Unicode fonts, custom typography and native release verification remain outside this contract.
VIEW-001
Local file bounds
The shared viewer currently accepts PDFs up to 50 MB and 500 pages. The Windows bridge streams in chunks, but the PDF engine still assembles a final contiguous buffer.
VIEW-003
Viewing modes
Zoom, fit, navigation, permanent page rotation, and an app-window presentation view exist. Presentation hides editing chrome and supports visible controls plus Arrow, Page, Home, End, Space, and Escape keys. It does not request browser or operating-system fullscreen, add transitions or timing, provide presenter or multi-monitor tools, or edit the PDF.
PRINT-001
Standard printing
A local raster preview supports all pages or one comma/range selection at 144 or 216 px/in before opening the browser or Windows print dialog. One preview is bounded to 50 pages, 20 megapixels per page, and 50 megapixels retained in total. Text, links, forms, tags, vectors, colour profiles, and measurement guides are not retained. Printer options and mixed-page scaling belong to the native dialog. Physical printing and Windows print-to-PDF output are not yet fixture-verified; inspect every native preview.
ORG-001
Complex structures
Organizer export stops on forms, bookmarks, signatures, encryption, tags, attachments, XFA, active content, earlier revisions, and damaged structure when fidelity cannot be established. Page content, resources, annotations, boxes, and rotation are copied into a new PDF, but document-level metadata is not copied.
CREATE-001
Blank PDF creation
Creates 1–500 identical local blank A4, US Letter, or US Legal pages in portrait or landscape. The selected 0, 36, or 72 point content-safe margin is stored as the standard ArtBox on every page; ordinary readers may not display that guide. The preview represents one page’s geometry. Custom sizes, mixed sizes within one creation, templates, visible guides, and automatic editor opening are not implemented.
ORG-008
Image insertion
Still PNG, JPEG, and WebP images are supported locally up to 50 MB and 20 megapixels each. EXIF orientation is applied before page creation. WebP and oriented images are normalized to PNG, so their image metadata is not copied. Animated WebP stops safely; choose a still frame first.
ORG-009 · ORG-010
Crop and resize
Crop changes the visible boundary only; hidden content remains recoverable and it is not redaction. Resize supports one range with fit/fill, but blocks forms, annotations, tags, print-production boxes or metadata, unusual page units, and risky structures. Multiple ranges, direct rendered-page drag handles, and annotation-aware resize are not implemented.
ORG-011 · ORG-012
Page and Bates numbering
One local range supports basic-Latin visible identifiers. Alphabetic labels use A–Z, AA–ZZ, AAA–ZZZ, not spreadsheet columns. Spaces are literal; control characters are rejected. Identifiers are limited to 120 characters and must fit the page. Unsupported characters or geometry stop export without a partial PDF or report. The sequence preview does not check placement, overlap, or the final count when the last page is unknown. Inspect the exported copy. Page-label mode replaces existing labels only after confirmation; Bates mode preserves them. Multiple ranges, custom fonts, conflict detection, removal, batch folders, and evidence or chain-of-custody claims are not supported.
EDIT-002
Text changes
Add text creates a new overlay; Edit existing text changes supported original string operators using the original font. It does not edit text inside form XObjects, substitute fonts, add missing glyphs or reflow paragraphs. OCR leaves scanned words visually unchanged. Both paths require review of the actual output and preserve originals with undo.
EDIT-006
Headers and footers
Basic-Latin text supports page tokens, alignment, size, margins, and custom ranges such as 1,3-5. Review the actual generated PDF before downloading; a settings or file change clears the preview. Unsupported characters, clipped headers, and unusual page geometry stop safely. Advisory checks distinguish ink near estimated text areas from ink elsewhere in full-width margin bands on visited selected pages only. Estimates include spaces and guards, not exact letter shapes. Backgrounds can trigger warnings; faint, tiny, hidden, or unrendered content may be missed. Missing metrics retain the broad warning; oversized or unreadable pages show unavailable, never safe. Always inspect placement manually. Exact overlap detection, custom fonts, and removal are not implemented.
EDIT-007
Watermarks and backgrounds
Visible text watermarks share custom ranges and the actual-PDF preview, with bounded opacity and angle controls. Margin sampling does not check watermarks; inspect their placement manually. Background replacement, custom fonts, images, background layering, and removal are not implemented. Encrypted, signed, XFA, attached-file, and active-content PDFs stop safely.
ANNO-003
Eraser
Freehand strokes support whole-mark deletion and bounded partial erasing by pointer or keyboard. The path is resampled and retained segments become separate marks, so inspect the exported copy. Pointer pressure is not retained or claimed.
ANNO-006
Stamps
Built-in and custom basic-Latin text stamps flatten into a new copy. Optional saving is device-only and bounded to 12 records. Stamps do not prove identity, approval, signature, workflow state, or evidence. Image/dynamic stamps, encrypted asset libraries, import/export, and team sharing are not implemented.
ANNO-007
Comment reviews
The review ledger covers PDFrig-created local comments, open/resolved status, page/status filters, and a content-bearing JSON summary. It does not import native PDF comments or provide replies, verified identity, assignments, mentions, collaboration, sync, notifications, shared links, or audit/evidence proof.
ANNO-008
Measurement guides
One guide per page supports calibrated straight-line distance or rectangular area/perimeter. A known reference and uniform scale are required. Precision is display resolution, not accuracy. Guides are session-only and omitted from PDF export/printing. Polygons, polylines, snapping, native measurement objects, persistence, reports, and survey/construction accuracy claims are not supported.
FORM-001
Forms
Common AcroForm text, checkbox, radio, and dropdown fields are supported. XFA and push-button actions are not.
SIGN-001
Visual signatures
Typed, drawn, and uploaded marks are visual signatures only. They do not validate identity or create a certificate-backed digital signature.
OPT-001
Compression
The current local optimizer is lossless structural optimization. It does not promise a smaller file or image-quality presets yet.
OCR-001
OCR beta
English and French worker fixtures pass locally. Production upload orchestration, quality benchmarks, cost evidence, legal review, and independent security review remain incomplete.
REDACT-001
Redaction beta
The current method rebuilds the whole document as verified image-only pages, removing search, forms, tags, links, vectors, and signatures. It stays disabled pending broader corpus and independent review.
