Choose supported file formats for selectable text
Choose a format that preserves the needed text
The supported file formats guide separates extraction support from a filename extension. PDF needs a selectable text layer, DOCX supplies main document text, and TXT is decoded as UTF-8. Choose the source that best preserves the passage you need to count. A scanned PDF may look readable but still contain only images. A Word document saved as DOCX may be easier to inspect before export. The supported file formats list does not imply that every feature of an office document, archive or ebook is interpreted or preserved.
Convert deliberately and inspect the result
When a format is not supported, export a copy from a trusted editor instead of merely changing its suffix. Inspect the beginning, middle and end after conversion, including accented characters and table text. The plain-text tool may accept a Markdown or CSV file as raw text, but that is not dedicated Markdown rendering or column-aware CSV parsing. The supported file formats table distinguishes those cases from planned support. For mixed sources, use the batch tool and check every row. Keep the original documents so a conversion problem can be traced and corrected.
Check extraction quality in supported file formats
Set the scope before checking a limit
For an essay, count the text your school actually includes in the limit. A title page, bibliography and figure captions may follow different instructions. Paste the main argument first, then count any additional sections separately. When editing a long article, compare the introduction and conclusion with the central sections. A total can show that a draft is too long, but it cannot tell you which argument needs more evidence. Keep a copy in your editor so you can revise and compare the same passage again.
Understand punctuation and mixed languages
The counter separates ordinary words at whitespace. A contraction such as don’t and a hyphenated expression such as well-known each remain one word. A number such as 3.50 also counts as one. Punctuation by itself adds no words. Chinese characters and Japanese kana are counted individually, including next to Latin text without spaces. This is a reproducible counting convention, not linguistic word segmentation. For a bilingual assignment, check whether the recipient wants words, Chinese characters or a separate total for each language before interpreting the result. Character totals use UTF-16 units rather than visible symbols. An emoji or a letter with a separate accent can occupy several units. For a strict form limit, check the destination’s own counter as well, because its counting convention may differ.
Compare the same passage in Word and Google Docs
If another editor gives a different result, first select exactly the same passage in both tools. Copying a whole document can introduce repeated headers, page numbers or line breaks that were absent from the original selection. A DOCX file is counted from its main document text; separate footnote and header files are not read. A PDF depends on its extractable text layer and may contain awkward line endings. Inspect the extracted preview before relying on the total. Then compare a short sample containing the punctuation or mixed writing that caused the difference.
Separate text privacy from browser storage
Counting runs on your device without sending the pasted text or document contents to a counting server. ZipCount does not save the draft for your next visit, so keep your own source file before closing the page. Local processing does not control your clipboard, browser extensions or other people using a shared computer. Avoid leaving a private draft visible after you finish. For a speech, keep a separate rehearsal copy: the estimated duration is a starting point, and a timed reading with pauses gives a more useful delivery plan.
Keep a useful record
Before you start, keep the source document and write down what the receiving editor or submission form includes. After checking the result, note the filename, language, selected passage and counting convention alongside the total. A number copied without that context is difficult to compare later. If you revise the source, repeat the check on the revised version rather than combining totals from different drafts. Browser memory is finite: if a large file stalls, close other heavy tabs and try a smaller document. There is no promised maximum file size. A failed extraction is not evidence that the original document contains no words.