Appendix C — App cheat sheets: one printable card per app
Summary
One compact card for every live app, in book order. Each card has the same six lines: purpose, input, output, key options, works before/after, gotcha. Print the page (each card is a short block) or copy the ones you use. Not-available apps (Redactor, Compare PDFs) and retired apps (Release Notes, Fork Sync, Personal Engineer) are listed at the end, not carded. Details and "verify on the live app" caveats are in each app's chapter.
Details
Part IV — Organize
Merger — chapter
- Purpose: many files into one PDF, in file order.
- Input: PDF, Office, HTML, images, DJVU, MOBI, archives (encrypted files refused).
- Output: one PDF.
- Key options: Mode (Sequential / Vertical / Horizontal / Grid), Landscape, Page size, OCR, DPI.
- Before / after: Rotation, Deskew, Unlock / Add TOC, Resize, Compressor.
- Gotcha: orders files, not pages. Site FAQ: up to 10 files and about 30 MB per operation.
Splitter — chapter
- Purpose: cut one PDF into pieces or pull out pages.
- Input: PDF. Output: several PDFs (or one with "range into a single file").
- Key options: six methods (every page, odd/even, every N, by page number, ranges to many files, ranges to one file).
- Before / after: Unlock / Merger, Rotation, Compressor, Protector.
- Gotcha: physical page numbers, not printed ones; pieces together can exceed the original size.
Delete Page — chapter
- Purpose: remove pages. Input: PDF. Output: PDF without them.
- Key option: page list such as
2,6-10. - Before / after: Search (find pages) / Add TOC, Compressor.
- Gotcha: the old table of contents may now be wrong.
Rotation — chapter
- Purpose: turn pages 90/180/270 or auto-straighten. Input: PDF. Output: PDF.
- Key options: Scope (all/selected), Angle (incl. Auto Skew), Output orientation, Document language.
- Before / after: before any OCR; around Merger / Searchable.
- Gotcha: Auto Skew re-renders pages; for crooked scans use Deskew.
Resize — chapter
- Purpose: change paper size or tile several pages per sheet. Input: PDF. Output: PDF.
- Key options: Page size, Tile mode, Pages per sheet (2/4/6/9/16), OCR + languages (tile mode).
- Before / after: Merger, Rotation / Booklet, Compressor.
- Gotcha: tile mode turns pages into images; tick OCR to keep text searchable.
Booklet — chapter
- Purpose: reorder pages for folded printing. Input: PDF. Output: one PDF (Auto) or two (Manual duplex).
- Key options: Layout (saddle-stitch / accordion), Panels per sheet, Duplex, Orientation, Page rotation.
- Before / after: Resize (half-sheet source), Delete Page, Add TOC / print.
- Gotcha: blank pages are added to reach a multiple of 4; that is expected.
Part V — Add & Structure
Add TOC — chapter
- Purpose: write your own contents page and bookmarks. Input: PDF. Output:
_with_toc.pdf. - Key options: TOC Title, bookmarks on/off, entries (text, page, level, numbering).
- Before / after: Merger, Delete Page, Splitter, Flatten / Compressor, Protector.
- Gotcha: trusts your page numbers; whether numbers count the inserted contents page is unclear (verify on the live app).
Table of Contents AI — chapter
- Purpose: detect headings, build a bookmark outline. Input: text PDF. Output: PDF with outline.
- Key options: Max Layers and "Only by extracting knowledge" (may change nothing; verify).
- Before / after: Searchable, Merger / Compressor, Protector.
- Gotcha: AI result varies between runs; skim the outline. Free plan may stop early with a continuation.
Add Table — chapter
- Purpose: insert Excel data as table pages. Input: PDF +
.xlsx. Output:_with_table.pdf. - Key options: Insert Page Before, Sheet Name, header row, Maximum Rows/Columns.
- Before / after: Merger / Delete Page; Add TOC, Compressor.
- Gotcha: CSV and old
.xlsfail ("no XLSX found"); only one sheet per workbook is read.
Flatten — chapter
- Purpose: freeze form fields into page content. Input: PDF. Output:
_flattened.pdf. - Key options: Fields to Skip, or Fields to Flatten (use one).
- Before / after: Extract Form Data (save answers first) / Add TOC, Sign, Compressor.
- Gotcha: irreversible on that copy; mistyped field names are silently ignored.
Collage — chapter
- Purpose: AI-laid-out photo sheets. Input: images (JPG/PNG/WebP/BMP; FAQ: up to 30). Output:
collage.pdf+ preview. - Key option: Landscape (the only one on the page).
- Before / after: Deskew for scans / Add TOC, Compressor, Booklet.
- Gotcha: takes images, not PDFs; the site's themes/borders/filters were not confirmed in code.
AI Illustrator — chapter
- Purpose: cover, borders, ornaments, AI paintings for a story. Input: TXT or PDF. Output: decorated PDF.
- Key options: Decoration Style, Story Genre, Cover, Borders, Ornaments, AI Paintings.
- Before / after: Chat, Converter, Add TOC / TOC AI, Compressor.
- Gotcha: generated art differs each run; at most 5 paintings per book; check the cover title/author.
Part VI — Protect & Prove
Protector — chapter
- Purpose: password-protect a PDF (and handle layers). Input: PDF. Output: protected PDF / ZIP.
- Key options: Password, Full Document Lock, layer lock/merge/extract/flatten.
- Before / after: all edits, Unlock / nothing (do not edit afterwards).
- Gotcha: the password cannot be recovered; one document-level password protects everything, even in layer modes.
Unlock — chapter
- Purpose: remove a password you know. Input: PDF + password. Output: PDF (ZIP for several).
- Key option: Password.
- Before / after: nothing / almost every other app.
- Gotcha: only for files you may unlock; unlocking can invalidate an existing signature.
Sign — chapter
- Purpose: digital signature with your
.pfx. Input: PDF + certificate + password. Output: signed PDF. - Key options: Certificate, Password, Mode.
- Before / after: everything else, signing is last / Verify Signature.
- Gotcha: signature box is fixed on the first page; no timestamp; a self-made certificate shows as untrusted.
Verify Signature — chapter
- Purpose: check signatures are unchanged. Input: PDF (one). Output: verdict on the page.
- Key option: Mode: Verify.
- Before / after: Sign / usually last.
- Gotcha: "valid" means unchanged, not "trusted" or "signer identified". Do not re-save first.
PDF/A Validator — chapter
- Purpose: check archive/accessibility compliance. Input: PDF. Output: on-screen report + JSON.
- Key options: Standard to validate against, Group identical problems.
- Before / after: Flatten, Optimize, Compressor (validate after them) / your PDF/A producing step.
- Gotcha: "not valid" is common (unembedded fonts, no output intent, encryption); confirm legally critical cases with the receiver's validator.
Part VII — Shrink, Clean & Repair
Compressor — chapter
- Purpose: smaller files by downsampling pictures. Input: PDF (FAQ: a single PDF up to 32 MB). Output: PDF + size report.
- Key option: Low (150 DPI) / Medium (120) / High (96).
- Before / after: edits, Deskew, Searchable / Protector, Sign.
- Gotcha: compress last, once; may save little on text-only PDFs.
Optimize — chapter
- Purpose: tidy internals, web-fast view. Input: PDF. Output:
_optimized.pdf. - Options: none.
- Before / after: Merger, Recovery, Flatten / Compressor, Protector.
- Gotcha: exact clean-ups are not itemised; savings can be tiny.
Deskew — chapter
- Purpose: straighten crooked scans. Input: PDF. Output:
Deskew_<name>.pdf. - Options: none (auto angle).
- Before / after: Rotation, Merger, Recovery / Searchable, Resize, Compressor.
- Gotcha: pages are rebuilt as images; do not count on a text layer surviving; run Searchable afterwards.
Recovery — chapter
- Purpose: repair damaged PDFs and sanitize. Input: one PDF. Output:
repaired.pdf+ report. - Key options: Expected page count, Minimum text coverage, Sanitization policy (default Full), X-ray report.
- Before / after: nothing / Compressor, Optimize, Searchable, Protector, Sign.
- Gotcha: default policy strips metadata and attachments; reconstructed pages may differ from the original.
Unwatermark — chapter
- Purpose: remove hidden AI-text watermarks (not visible stamps). Input: one text file or text PDF. Output: cleaned text + new PDF + report.
- Key options: Rewrite with AI (on), Detect-only, Threshold 35, strip vendor names.
- Before / after: Searchable, Converter / Converter, Add TOC.
- Gotcha: PDF layout is lost; rewriting can change meaning; detection is probabilistic.
Part VIII — Extract, Read & Search
OCR Image — chapter
- Purpose: text from pictures. Input: JPG/PNG (FAQ: up to 100 files, about 32 MB). Output:
.txtper image / ZIP / one file. - Key options: Document language, single output file.
- Before / after: Rotation, Deskew / your editor, Merger + Searchable.
- Gotcha: wrong language is the top cause of garbage; no layout in the output.
Searchable — chapter
- Purpose: add an invisible text layer to scans. Input: scanned PDF. Output: PDF (or DOCX/XLSX/PPTX).
- Key options: Language, Deskew, rotate 90, Remove background, AI correction, Output format.
- Before / after: Rotation, Deskew, Unwatermark, Unlock / Search, Extract Text, Add TOC, Ask PDF, Compressor.
- Gotcha: recognition is an estimate; AI correction can change words; file grows.
Extract Text — chapter
- Purpose: plain text out of a PDF. Input: text PDF. Output:
.txtper PDF (+ combined file). - Key option: Pure / Raw / Flatten.
- Before / after: Searchable, Unlock, Splitter / your editor, Ask PDF.
- Gotcha: text in images is not found; layout is dropped.
Extract Image — chapter
- Purpose: save embedded pictures. Input: PDF. Output: PNG files.
- Options: none.
- Before / after: Unlock, Splitter / OCR Image, Merger.
- Gotcha: a scan gives one big image per page; PNG may be larger than the original JPEG.
Extract Metadata — chapter
- Purpose: read document properties. Input: PDF. Output: table +
extracted_metadata.csv. - Options: none.
- Before / after: Unlock, Recovery / protect family, PDF/A Validator.
- Gotcha: values may be wrong or template names; no modification date in the app's list.
Extract Form Data — chapter
- Purpose: form answers to CSV. Input: filled fillable PDFs. Output: CSV per form + two summaries.
- Options: none.
- Before / after: Unlock, Splitter / spreadsheet, Flatten.
- Gotcha: empty or flattened forms give empty values; mix only the same template.
Parser — chapter
- Purpose: take a PDF apart (text, images, attachments, annotations, bookmarks, form data). Input: PDF. Output: ZIP.
- Options: none.
- Before / after: Unlock, Searchable / Extract Form Data, Extract Metadata, OCR Image.
- Gotcha: many files; scans give an empty text file.
Search — chapter
- Purpose: find words or patterns in your PDFs. Input: PDF. Output: PDF with highlights and a results bookmark tree.
- Key options: Search text, regular expression, scanned, language.
- Before / after: Searchable, Unlock, Rotation / Splitter, Compressor.
- Gotcha: only files with matches come back; on scans "no results" does not prove absence.
Global Search — chapter
- Purpose: find Documentize pages and articles. Input: typed words. Output: list of site pages.
- Gotcha: it never searches your documents; a third-party search box is used.
Resume Analyzer — chapter
- Purpose: AI analysis of CVs against a job. Input: PDF or DOC. Output: PDF report.
- Key options: Operation, Job title/description, Industry, Focus areas, Theme, report parts.
- Before / after: Searchable, Unlock, Merger / Compressor.
- Gotcha: AI opinion, not measurement; two or more CVs always run a comparison; the CV goes to an AI service.
Part IX — Convert & Create
Converter — chapter
- Purpose: format to format, in and out of PDF. Input: dozens of formats and archives. Output: target format (ZIP for several).
- Key options: OCR, Deskew, Page size, Landscape, Multi-layer, Excel mode, Delimiter, XML format, Base64.
- Before / after: Unlock, Rotation, Deskew / Merger, Compressor, Searchable, PDF/A Validator.
- Gotcha: content must match the chosen type; Office pagination may differ; FAQ: up to 10 files and about 30 MB.
Chat (create a PDF) — chapter
- Purpose: describe a document, get a PDF, revise by chat. Input: a request; optional PDF; up to 3 public links. Output: PDF.
- Key options: Message, Pages (1–50), Attach PDF.
- Before / after: Searchable, Extract Text / TOC AI, Illustrator, Converter.
- Gotcha: text only, AI may invent details; edit mode works up to 200 pages.
SVG Maker — chapter
- Purpose: chat to a layered vector drawing. Input: an idea; optional image/SVG. Output: SVG + PDF.
- Key options: Drawing style, Iterations (1–5, default 3).
- Before / after: nothing required / Converter, Merger.
- Gotcha: non-deterministic; the PDF is a picture, keep the SVG to edit.
Part X — AI: Understand
Abstracts — chapter
- Purpose: an abstract page inside your PDF. Input: PDF(s). Output:
_abstract.pdf+ text. - Key options: Style, Length, Language, keywords, conclusion, Insert after page, Document type.
- Before / after: Searchable, Merger / Add TOC, Ask PDF.
- Gotcha: AI; long documents are read in parts, free runs may sample.
Ask PDF — chapter
- Purpose: questions with page citations. Input: text PDFs (up to 10 files, 100 MB per app text). Output: cited answer.
- Key options: Answer length, Link answers to source pages.
- Before / after: Searchable, Deskew, Merger / Checklist, Cheat Sheet.
- Gotcha: images are not read; documents kept about a day; open the citation.
Cheat Sheet — chapter
- Purpose: dense one-page digest. Input: text PDF(s) or a topic. Output:
_cheat_sheetPDF. - Key options: Topic, categories, Density, Page size, Language.
- Before / after: Searchable, Splitter / Resize, print.
- Gotcha: topic-only content is unchecked model knowledge.
Checklist — chapter
- Purpose: prioritised, page-referenced checklist. Input: text PDF + required topic. Output: checklist PDF + summary.
- Key options: Topic, Type, Detail, Max items.
- Before / after: Searchable, Splitter / Ask PDF.
- Gotcha: page dropdown labels differ from engine values (verify); higher detail adds inference.
PDF Knowledge Base — chapter
- Purpose: wiki + chat over many PDFs. Input: text PDFs (up to 10 files, 100 MB per upload). Output: wiki,
pdfkb_wiki.zip, cited chat. - Key options: Summaries, Concepts, Entities, Explorations, Wiki language.
- Before / after: Searchable, Merger / Checklist.
- Gotcha: kept about a day (download the zip); long documents are sampled for summaries.
Part XI — AI for developers
Assembly Chat — chapter
- Purpose: scan, ask, stats, diff of a .NET assembly. Input:
.dll/.exeor NuGet name. Output: diagrams, CSV, answers, reports. - Key options: Command, Query, NuGet package/version, After package, Include Internal.
- Before / after: NuGet Updater / Repo Docs.
- Gotcha: only the public surface is known; compile the samples.
Code Graph — chapter
- Purpose: SQLite map of a C# solution. Input: GitHub repo with
.sln. Output:.sqlite. - Key options: Repository, Branch.
- Before / after: nothing / Repo Docs, Refactor Assistant.
- Gotcha: private repo needs
arise-projectinvited; build-time generated code is absent.
Repo Docs — chapter
- Purpose: documentation site as a pull request + zip. Input: GitHub repo. Output: PR, Markdown/HTML/PDF zip.
- Key options: Solution, Sections, Audience, Detail level, Output language, Docs folder.
- Before / after: Code Graph, NuGet Updater / MkDocs or GitHub Pages.
- Gotcha: free run stops after 15 minutes; code excerpts go to a language model.
Refactor Assistant — chapter
- Purpose: find oversized classes; optionally extract into a PR. Input: GitHub C# repo. Output: ranked report; PR.
- Key options: Branch, Candidates, Inspect one class, Verdict/Project filters, Apply mode.
- Before / after: Code Graph / Repo Docs.
- Gotcha: Apply needs sign-in and a paid plan, 10 to 30 minutes per seam, at most 5 seams; code is not compiled by Documentize.
NuGet Updater — chapter
- Purpose: report or bump outdated packages. Input: GitHub .NET repo. Output:
nuget-updates.md; PR. - Key options: Solution, Mode, Smallest update, Maximum packages, Verify by.
- Before / after: Assembly Chat / Repo Docs.
- Gotcha: "Watch weekly" is not scheduled yet; a PR may be a draft "not build-verified".
Page SEO Checker — chapter
- Purpose: audit H1/title/meta description, AI rewrites. Input: page or sitemap URL. Output: CSVs + JSON.
- Key options: What to check, Pages to check (default 25), texts, Target keyword, AI on/off.
- Before / after: Repo Docs.
- Gotcha: run stops at 15 minutes (free) or 30 (paid); login-protected pages are unreadable.
Advisor / Document Engineer — chapter
- Purpose: advice on which tool to use / a buildable .NET 10 project. Input: chat message. Output: answer / zip.
- Key options: answer style (Detailed, Concise, Code-first); Target OS.
- Gotcha: generated projects are built, never run; do not paste secrets.
Not carded
- Not available: PDF Redactor, Compare PDFs (the chapters give a manual route).
- Retired: Release Notes, Fork Sync, Personal Engineer (see advisors).