File processing & limits
Know what the current tools do, which files they accept, and what to check before saving your result.
How the current tools handle files
The eleven implemented tools listed below read and process your selected files in your browser. Their current processing paths do not send file bytes to a backend service. PDF operations use a browser Worker; Image to PDF also uses browser image decoding and, when needed, a local image/canvas compatibility path before Worker PDF assembly. PDF text and image exports use a local PDF.js Worker, bounded text/DOCX packaging and sequential canvas rendering.
This describes these tools, not all website or hosting activity. Loading the website requests pages and application code. Deployment logging, cookies and third-party hosting behavior have not been verified here; this guide makes no site-wide logging or cookie guarantee.
Preview tools do not process files
The following catalogue entries remain previews. Their processing action is disabled and they produce no converted output:
Current tool limits
MiB means 1,048,576 bytes. The 25 MiB per-file cap is 26,214,400 bytes; a file exactly at the cap is allowed by the size check, but must still pass format and content validation.
Generated PDF output is limited to 75 MiB. Worker tasks have a 60-second deadline; this is not a promise that the entire selection, inspection and download workflow finishes in that time.
PDF to Word
1 PDF, up to 25 MiB and 500 pages. Selectable text only, with 256 KiB text per page and 4 MiB combined. Word infers lines, simple label/value tables and clear columns from coordinates; retains page breaks, sizes and available bold/italic or underline-annotation metadata. Unsupported or ambiguous geometry falls back to text. No OCR, images, embedded original fonts, complex-table or exact-layout guarantee. The DOCX cap is 16 MiB. Check reading order, layout and pagination in your editor. Independent Word/editor compatibility is not guaranteed.
PDF to Markdown
1 PDF, up to 25 MiB and 500 pages. Selectable text only: 256 KiB per page and 4 MiB combined. Literal text is escaped into Markdown with page headings; no AI, OCR, semantic structure, images or table reconstruction.
PDF to JPG
One PDF up to 25 MiB, at most 50 pages. Render at 150 DPI with a white background: 4096 pixels per side, 12 million pixels per page and 36 million pixels combined. One page downloads as JPG, multiple pages as ZIP. Twice retained JPEG bytes plus 64 KiB ZIP reserve must fit the 75 MiB budget. Review complex fonts and images in the output.
Merge PDF
2–20 PDFs, 25 MiB per file, 50 MiB combined and 500 pages combined. Output follows the selected file order.
Split PDF
1 PDF, up to 25 MiB and 500 pages. Up to 50 output PDFs and 500 output pages combined. Page expressions are limited to 4000 characters. Multiple outputs also use the ZIP budget described below.
Rotate PDF
1 PDF, up to 25 MiB and 500 pages. Rotate all or selected pages by 90° clockwise, 90° counterclockwise or 180°. Page selections must be valid and must not repeat pages.
Organize PDF
1 PDF, up to 25 MiB and 500 pages. Keep at least one page. Pages may be reordered or removed; each retained page appears once. Keeping only the required pages also extracts them into a new PDF.
Compress PDF
1 PDF, up to 25 MiB and 500 pages. Lossless stream compression has Low, Balanced and High settings. Decoding is bounded to 8 MiB per stream, 32 MiB combined and 2000 streams. Unsupported or oversized streams are retained; malformed eligible streams can fail processing. Images are not resampled, and output may be larger: compare the actual sizes.
Add Page Numbers
1 PDF, up to 25 MiB and 500 pages. Starting and final numbers must be 1–999999; font size is 8–32 points. Selected pages must not repeat. Labels must fit the chosen position on the page.
Watermark PDF
1 PDF, up to 25 MiB and 500 pages. Nonempty single-line text, at most 100 characters, supported by the built-in font. Font size 8–96 points, opacity 10–100%, angle −180° to 180°. Selected pages must not repeat, and text must fit the page.
Image to PDF
JPG/JPEG, PNG and static WebP: up to 20 images, 25 MiB each and 50 MiB combined. Each image is limited to 8192 pixels on either side and 12,000,000 pixels; the set is limited to 36,000,000 pixels. Each image becomes one PDF page in the selected order.
Split output and ZIP budget
For multiple outputs, twice the combined PDF output bytes plus a 64 KiB ZIP reserve must fit within 75 MiB. This can reject a split even when its PDFs alone are below that cap. The completed ZIP must also be no larger than 75 MiB. A single output uses the PDF byte budget without that ZIP reserve.
Image page settings and conversion budget
Choose A4, Letter or image-sized pages. Portrait and landscape apply to A4 and Letter. Image-sized pages use 96 pixels per inch, plus margins. Margins range from 0 to 30 mm. Contain keeps the whole image; Cover crops to fill the area inside the margins.
Converted image data must also remain within 25 MiB per image and 50 MiB combined. Conversion can increase size, so accepted originals can still exceed this later budget. Images are decoded sequentially during conversion to reduce simultaneous allocation.
Before selecting or processing a file
- Keep your original files. PDF tools require readable, unencrypted PDFs; password-protected PDFs are unsupported. Empty, corrupted, unsupported or excessive inputs can be rejected during inspection or processing, even after selection.
- Image to PDF does not accept HEIC, GIF, SVG or animated images. Orientation depends on browser decoding support. Original image metadata and exact color appearance are not guaranteed to be preserved.
- Changes may invalidate PDF digital signatures. Forms, bookmarks, attachments and other document features are not guaranteed to survive every operation.
- Watermarks and page numbers are visible overlays. They do not redact underlying content or password-protect a file. These tools are not a general PDF sanitization service.
- Byte and page caps reduce resource exposure; they do not guarantee a safe peak memory requirement. A small compressed file can expand substantially. Large or complex files may fail on devices with limited resources.
Downloads, cancellation and recovery
- Wait for completion before downloading. Output is checked structurally by the processing code, but this is not a visual fidelity certification. Open the downloaded PDF in your usual viewer and check its pages, order, text and appearance.
- Downloads use browser Blob URLs. Save the output before resetting or leaving the tool. The app does not provide a cloud copy or cross-session recovery.
- Use Cancel while processing, where offered. After an error, review its message, adjust the selection or settings, and retry. Cancellation during image inspection may require selecting unfinished images again.
- Reset starts a new operation and releases workspace resources and download URLs. It does not delete original files or saved downloads, guarantee immediate memory reclamation, or securely erase browser memory. Reloading or navigating away can lose the current selection and result.
Browser and device limitations
Processing needs browser Workers and download support; image conversion also needs supported image decoders and browser image/canvas APIs. Available memory, background-tab suspension and native file pickers can affect completion. Keep the processing tab open and active.
Firefox, native Safari, physical phones/tablets, manual screen-reader behavior and browser zoom still require further verification. Automated desktop or emulated-mobile checks do not establish support on every device. Exact Worker and native peak memory is not yet measured. Deployment behavior is also unverified; the project is not release-ready.