Write your post, add bold, italic and lists, see where the feed cuts it off, and copy it so the paragraphs survive.
Formatting uses Unicode lookalike characters. Screen readers and LinkedIn search may not read them like normal text.
Blank lines are kept when you paste into LinkedIn.
Without a selection, the word under the cursor is formatted.
Applying a style replaces the previous one. Applying the same style again removes it.
Digits stay normal in italic and bold italic; Unicode has no matching digits.
Hashtags, mentions and links should stay unstyled so they remain clickable.
3,000 characters per post, counted like the copied text.
Mobile preview: cut at about 140 characters or the third line break.
Desktop preview: cut at about 210 characters or the third line break.
The cut-off values are observed approximations and can change.
LinkedIn has no formatting buttons in normal posts. Bold and italic text there is made of Unicode lookalike characters, and blank lines only survive pasting with a small workaround. Here is what works and what to watch out for.
LinkedIn's post editor has no bold or italic button. Tools like this one replace letters with look-alike characters from the Unicode block “Mathematical Alphanumeric Symbols”. They look bold or italic everywhere, because they are different characters, not a font setting.
Select a word or sentence and choose a style, or press Ctrl/Cmd + B. Use it for the opening line or a few key phrases, not for whole paragraphs.
In effect, yes, but no font is involved. LinkedIn always shows its own typeface. What looks like a different font is a set of Unicode characters: sans-serif bold, italic, bold italic, serif bold and monospace, plus underline and strikethrough made with combining marks.
Because they are characters, the styles survive copy and paste and look the same on LinkedIn, in the app and in comments. Ordinary letters, umlauts and emojis stay as they are, and “Remove style” turns styled text back into plain text.
Paste text from ChatGPT, Notion or another Markdown editor and choose “Convert Markdown”. Headings and **bold** become Unicode bold, *italic* becomes italic, `code` becomes monospace, list markers become “•” or “1.”, and links become “text: URL”.
Code blocks and URLs are protected, so underscores in snake_case names or links stay intact. Undo restores the original if you want to compare.
When you paste text, LinkedIn often merges empty lines, so your paragraphs run together. When you copy, every empty line is filled with the Braille blank U+2800. It looks empty, but the editor treats it as content and keeps the gap.
The copy button also puts each line into its own paragraph in the HTML version of the clipboard. Both together worked in our tests; LinkedIn can change this behaviour at any time.
Styled letters are separate characters. Some screen readers read them letter by letter or announce them as mathematical symbols, and search may not match “𝗟𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽” to “leadership”.
Keep important terms and names in plain text and use styling sparingly, so your post stays readable and findable.
LinkedIn only links hashtags and mentions written with normal characters. “#𝗠𝗮𝗿𝗸𝗲𝘁𝗶𝗻𝗴” is just text and does not lead to the topic page. The hints next to the preview warn you when a hashtag or mention is styled.
Open this page in Safari or Chrome on your phone, write or paste your post and tap “Copy post”. Then switch to the LinkedIn app, start a post and paste. If your browser blocks copying, the text is selected for you; use the system “Copy” menu instead.
Unicode styles look the same in the app because they are characters. Whether blank lines survive can depend on the app version; check the post before publishing.
There are no real headings, links with custom text, colours or font sizes in normal posts. Underline and strikethrough use combining marks that can look uneven in some fonts. A post can have up to 3,000 characters, and the “…more” cut-off depends on device and layout.
This tool shows observed approximations, not official LinkedIn rules. Check the result in LinkedIn before you publish.
The AI assistant works from your notes or your draft. It has a fixed list of phrases it must not use, such as “In today's fast-paced world”, “Moreover”, “game-changer” or a tacked-on “Thoughts?”. On top of that come rules you notice when you read a post out loud: sentences of different lengths, active voice, plain words, at most one dash, and no paragraphs that are all built the same way.
After every answer the server checks the text against that list again. If something slipped through, it asks once for a rewrite of exactly those spots. Anything left is shown as a note above the suggestion.
No, it is told not to. It takes concrete details, numbers and stories only from what you enter. If something is missing that would make the post stronger, it tells you in a note instead of inventing it. Those details are what makes a post credible: they show that someone is writing from experience, not about a topic in general.
Still, read the suggestion before you use it. It only goes into the editor when you click “Apply”, and “Undo” brings your old text back.
The topic and the main point go into the first one or two lines, before “…more”, using the term someone would actually search for. One post has one focus and your own view on it. A general summary that already exists a hundred times helps nobody.
Short paragraphs, lists only for real lists, bold only for the opening and one or two key terms. At most three specific hashtags at the end; you can switch each suggested one on or off. Instead of a generic closing question you get one that fits your topic.
As long as you don't start the assistant, your text stays in your browser. Only when you click “Create suggestion” is the text in the assistant sent to the configured AI provider. Which provider that is and what it stores is described in the privacy policy.