Text Tools

Counting, and why counters disagree

Word Counter updates five figures on every keystroke — words, characters, sentences, paragraphs and reading time — so you can trim toward a limit and watch the number move rather than pressing an Analyse button and waiting.

It states its rules instead of hiding them, which matters the moment an editor's count disagrees with yours. A word is whatever has whitespace on both sides after the text is trimmed, so a hyphenated compound counts once and a dash typed closed-up joins two words into one. A character is a UTF-16 code unit, which is why one emoji can add two to the count and a flag adds four. Sentences are found by splitting on runs of full stops, question marks and exclamation marks, so an abbreviation or a decimal point inflates the figure. Reading time is the word count divided by 200 and rounded up, a deliberately cautious pace. Knowing the rule lets you predict the gap when someone else's tool reports a different number.

Comparing two versions

Diff Viewer takes two blocks of text — drafts, config files, log excerpts, source code — and shows the changed lines side by side with the exact characters that differ marked inside them. That second level is the part that saves time: a line-only diff tells you a line moved, while a character-level diff tells you that one number in it changed.

It is the right tool whenever a document arrives back with "a few small edits" and you need to know precisely which. For comparing two PDF documents rather than two blocks of text, use Compare PDFs instead, which adds a pixel overlay so you can also see what moved on the page.

Rich text in, Markdown out, and back again

These three tools form a pipeline, and each is useful on its own.

Paste to Markdown takes whatever is on your clipboard from Google Docs, Word, Notion or a web page and converts the underlying HTML into GitHub-flavoured Markdown — with real pipe tables and fenced code blocks, not a flattened approximation. It is the fastest route out of a word processor and into anything that stores plain text.

Markdown Previewer renders Markdown live beside the editor as you type, so you can check that a table lines up or a nested list nests before committing it anywhere. It will export the rendered result as a standalone HTML file, and for a paginated document with page breaks and selectable text there is Markdown to PDF.

Humanize AI Text handles the punctuation that survives a copy and paste and then breaks something downstream: curly quotes are straightened, em and en dashes become plain hyphens, and zero-width characters — invisible in every editor, and quite capable of breaking a CSV import, a URL slug or a code comparison — are deleted outright. It is a character cleanup and nothing more; it does not rewrite your sentences.

Filling a layout before the copy exists

Lorem Ipsum Generator produces 1 to 100 paragraphs, sentences or words of the standard placeholder text, redrawn on every run. Designers reach for it because real copy is distracting during layout and because English placeholder text makes people start editing the words instead of judging the spacing.

Pair it with Word Counter when a component has a hard limit: generate to a length, count it, and you know what the real copy has to fit into before anyone writes a word of it.

What these tools deliberately do not do

None of them upload anything, so none of them can offer collaboration, saved documents or version history — refreshing the page clears the box and leaves nothing behind. None of them check spelling or grammar. And none of them rewrite meaning: the cleanup tool edits characters, the counter only counts, and the diff only reports. That narrowness is what makes their output predictable enough to trust with text you cannot send to a server.

For text that has to become a formatted document, cross over to the PDF family: Create PDF paginates plain text into a PDF with real selectable text, and HTML to PDF does the same for markup you already have.