Converter

End-user guide for the file converter side of frogConvert. For the PDF editor, see PDF_EDITOR.md. For programmatic access, see INTEGRATIONS.md.

Install

frogConvert installs as an app on Windows, macOS, iOS, Android, and any Chromium browser. Look for the install icon in your address bar (or "Add to Home Screen" on iOS).

Entry points and offline behavior

Once installed:

• Share sheet (Android). Long-press a file in any app and pick frogConvert from the share menu. The Converter opens with the file already loaded.
• Share sheet (iOS). Limited. iOS Safari does not fully support the Web Share Target API; depending on your iOS version the share menu may only launch frogConvert without auto-loading the file. Use the in-app upload zone as a fallback.
• "Open with…" On Windows, macOS, and ChromeOS, right-click a file and pick frogConvert. Registered for image, video, audio, PDF, text, ZIP, and 7z extensions.
• Offline. Conversion handlers (FFmpeg, ImageMagick, etc.) cache as you use them, so repeat conversions work without a network.
• Resume. Close the tab mid-conversion and frogConvert offers to Resume your work next time you visit. Files, target format, and selection survive the round-trip. Sessions older than 7 days are dropped automatically.

Converting a file

1. Upload - drag and drop a file onto the upload zone, or click it to browse. You can upload multiple files at once (up to the device limit).
2. Auto-detect - frogConvert automatically detects the input format and switches the category tab to match.
3. Pick an output - click the format selector to open the format picker. Browse by category tab or search.
4. Convert - hit Convert. A progress indicator shows how many files have been processed.
5. Download - converted files download automatically.

Tips

• Any-to-any. frogConvert chains multiple conversion tools to reach formats no single tool supports directly. WAV to PDF? Go for it.
• Theme toggle. Light/dark mode button in the top bar.
• Mode toggle (Core / Plus / All). Controls how many output formats are shown:
  • Core - common everyday formats only.
  • Plus - adds data, font, and extra media formats.
  • All - every supported format.
• Multiple files. Use the file manager to review, add, remove, or replace individual files.
• Partial downloads. Cancelled a batch mid-way? You can still download the files that finished.
• Compression while converting. Handlers like FFmpeg, ImageMagick and pdftoimg re-encode, so a conversion is also a chance to compress the file. Pick the level under Compression in the hamburger menu: Original quality (default - the format changes and nothing else), Automatic (reads each file and picks a tier), High quality, Balanced, or Smallest file. The same choice is available to agents as the quality argument on the MCP convert_file tool and the REST POST /convert endpoint. See INTEGRATIONS.md § Quality preset and COMPRESS.md.
• Picking the same format in and out does nothing, on purpose. The Converter changes the format; it does not re-encode a file into itself. Choosing JPG to JPG hands back your original bytes with a "No conversion needed" notice, and the button relabels to Download original so you know before you press it. The format picker signposts the alternative - "Same format in and out, so there's nothing to convert. Want it smaller?" - with an Open Compress button that carries the files across. Making a file smaller is Compress's job, and it is better at it: batches, PDFs, per-file reporting and the keep-threshold all live there. (Earlier builds did compress inline here, which is why the option existed at all.)
• Smart auto-adaptation. When a conversion would hit a browser-memory or sanity ceiling (very large PDFs, long videos to GIF, thousands of frames extracted from a long video), the pipeline adjusts the output instead of erroring. Adjustments are explained in a post-conversion notice card. See HANDLERS.md § Post-conversion notices for how handlers emit these.
• Performance. frogConvert detects available RAM and adjusts limits to prevent crashes on lower-end devices.

PostScript, EPS and Illustrator

Ghostscript - the same engine that powers PDF compression - reads the PostScript family natively, so these routes keep vector content as vector: text stays selectable and curves stay curves.

Route	Notes
PS → PDF, EPS → PDF, AI → PDF	The main direction. Once it is a PDF, everything else the app does with PDFs works on it: PNG/JPEG, text extraction, the PDF Editor, Compress.
PDF → PS	One PostScript file, all pages.
PDF → EPS	One file per page. An EPS cannot hold more than one page by definition, so a 10-page PDF gives you 10 files.
PDF → PDF/A	PDF/A-2b, for archival deposit. Anything that cannot be represented in the profile is dropped rather than failing the conversion.
PDF → TIFF	Multi-page, LZW-compressed. Resolution follows the compression level: 96 dpi at Smallest file, 150 at Balanced, 300 at High quality.

The compression level applies to these routes too. It controls how far embedded images are downsampled, so it does nothing to a purely vector file and a great deal to a scan. Measured on a 10 MB image-heavy source:

Route	Smallest file	Balanced	High quality
PS → PDF	128 KB	508 KB	1.07 MB
PDF → PDF/A	131 KB	511 KB	1.07 MB
PDF → PS	3.8 MB	16.3 MB	16.3 MB

Two things worth knowing:

• .ai files convert through their PDF layer. Illustrator has written PDF-compatible .ai since version 9 (2000), so the artwork comes across intact - but layers, editable text and effects are flattened. Keep the .ai as your master. The Converter says so before you press the button.
• The engine is a ~16 MB download, fetched the first time you use a PostScript route (or compress a PDF) and cached afterwards. It is never fetched at page load.

Known limitations

• PDF input on Safari. Safari's JavaScript engine cannot handle PDF input for conversions (PDF to PNG, PDF to TXT, etc.). Use Chrome or Firefox for PDF input. Other formats work normally on Safari.
• Office documents (DOCX, PPTX, XLSX, ODT, ODP, ODS) to PDF. A faithful PDF needs LibreOffice. The desktop app uses it when it is installed on the computer (restart the app after installing), and so do the command-line tools, MCP server and REST API (see INTEGRATIONS.md). In a desktop browser, the Converter offers the browser Office engine: LibreOffice compiled to WebAssembly, a one-time download of about 74 MB that stays on the device (Cache Storage, so clearing site data removes it). After it, DOCX, PPTX, XLSX, ODT, ODP and ODS become real PDFs with selectable text, a few seconds each, nothing uploaded. It is not offered on phones (it needs about 1.5 GB of memory) or in browsers without cross-origin isolation. Without it, and in the desktop app without LibreOffice, the PDF is a picture of the document: not selectable, and slide layouts and images may be lost. The Converter says so before you convert and again on the result. Legacy .doc and .ppt are not supported in the browser: the engine hangs on them, so it does not take them.
• HTML and Markdown to PNG or PDF. The page is laid out the way a browser shows it, with its own stylesheet, at a fixed width (1024 px for images, A4 for PDF) and drawn at twice that size, so the result does not depend on your window. A PDF is cut into A4 pages between lines of text. In a browser the PDF is pictures, not selectable text, and the result says so. For selectable text, the result offers Print as PDF: it opens the browser's print window on the page, already set to A4 with backgrounds on and no date or address stamped on the pages; choose Save as PDF and save. (A web page cannot make a PDF with text in it without that window; browsers do not allow it.) The desktop app prints it with Chromium's own PDF printer, so there the text is real. External stylesheets, web fonts and images are not fetched (converting never touches the network), so a page that relies on them looks plainer.
• Lossy routes are labelled. When the only route to a document format passes through an image (a PDF to DOCX or PPTX arrives as one picture per page) or through plain text, the result says so. See src/conversion/routeQuality.ts.
• Encrypted files are not supported. Remove the password first.

See also

• PDF_EDITOR.md - merge, reorder, extract, and watermark PDF pages.
• INTEGRATIONS.md - MCP and REST API for scripting conversions.
• ../SECURITY.md - privacy posture and threat model.