1. Overview
We need a desktop application (Windows & macOS required; Linux optional) that allows users to upload and edit any type of document:
- PDFs
- Images (JPG, PNG, TIFF, HEIC, WebP, etc.)
- Flyers, posters, cards, scanned documents
The app must include OCR (Optical Character Recognition) so text in scanned images becomes editable. Users should be able to edit text, replace images, annotate, redact, and export documents.
This tool should go beyond existing PDF editors like Sejda by combining:
- PDF editing
- Image editing
- OCR with layout analysis
- Non-destructive editing layers
- Advanced export options (PDF/A, PDF/X, CMYK, DOCX, image sets)
2. Goals
- Provide a desktop-first, offline-capable tool.
- Universal file import (PDFs, images, SVGs).
- Enable text and image editing directly in the document.
- Integrate OCR with multilingual support (English, Arabic, Chinese, Hindi, etc.).
- Export documents with print-grade quality (CMYK, ICC profiles, bleeds).
- Ensure performance and usability for large files.
3. Core Features (MVP Scope)
3.1 File Import / Export
- Import: PDF, PNG, JPG, TIFF, WebP, HEIC, SVG.
- Export: PDF (Standard, PDF/A, PDF/X), DOCX (basic text export), PNG/JPG, SVG.
3.2 OCR (Optical Character Recognition)
- OCR for multiple languages.
- Detects text, tables, and forms.
- Map recognized text back into editable layers.
3.3 Text Editing
- Edit existing text directly in PDFs or scans.
- Add/replace fonts (with fallback).
- Support for bold, italic, size, color, alignment.
- Spell check + search/replace.
3.4 Image Editing
- Replace or insert images.
- Crop, rotate, resize, deskew.
- Background removal (basic AI).
- Noise cleanup and resolution enhancement.
3.5 Layout & Objects
- Add/remove/reorder pages.
- Add shapes, lines, annotations.
- Align, snap, and layer controls.
- Group/ungroup objects.
3.6 Forms & Annotations
- Add/edit form fields (text fields, checkboxes, signatures).
- Comments, highlights, stamps.
3.7 Redaction
- True secure redaction (content permanently removed).
- Text and image redaction.
4. Advanced Features (Phase 2+)
- Batch OCR + conversions.
- AI helpers (“clean this scan,” “extract table to CSV,” “summarize text”).
- Plugin system for custom filters & imports.
- Pre-press features: overprint preview, bleed, preflight checks.
5. Performance Requirements
- Load a 300-page PDF within 3 seconds.
- Page render time <500ms.
- OCR accuracy: ≥98% on clean Latin text, ≥90% for complex scans.
- Crash-free sessions ≥99.9%.
6. Platform & Technology
We are open to the best stack. Options include:
- Qt/C++ (cross-platform, native performance)
- .NET (WPF/MAUI) (Windows-first, mac optional)
- Electron + Rust/C++ modules (fast UI dev, native processing)
Mandatory:
- Auto-update system (signed installers).
- Offline-first (OCR and editing must not require cloud).
- Modular architecture for future AI integration.
7. Deliverables
- MVP (3–4 months):
- Import/export
- OCR + editable text layers
- Text + image editing tools
- Page management
- Export to PDF/DOCX
- Beta Release (6 months):
- Forms & annotations
- Redaction
- Batch OCR
- Spell check + search/replace
- Packaging (Windows .exe, Mac .dmg)
- V1 Release (9–12 months):
- AI helpers (clean scans, background removal)
- Advanced export (PDF/A, PDF/X with CMYK)
- Plugin SDK
8. Success Metrics
- Users can open, OCR, edit, and export a PDF without leaving the app.
- Edited text is seamlessly integrated (fonts, flow preserved).
- Image edits (crop/cleanup) apply non-destructively.
- Export files open correctly in Adobe Acrobat, MS Word, and print workflows.
9. References
- Comparable Product: Sejda PDF Editor
- What We Want Beyond Sejda:
- OCR integration
- Mixed image + vector editing
- True offline-first desktop app
- Print-grade output
...
Show more