LinkedIn text formatter: Unicode and a readable layout
A LinkedIn text formatter does two jobs that people treat as one. The first is a letter swap: plain characters become lookalikes from the Unicode Mathematical Alphanumeric Symbols block, so a word looks bold or italic in a composer that has no style buttons. The second is layout: where the line breaks fall, which marker starts a list, and which characters are sitting in the draft that you cannot see. The letter swap is documented, with the code points and the missing italic h, on LinkedIn text formatting. This page is the layout. A styled post that cannot be scanned on a phone has failed the only test that matters.
LinkedIn's own help page says a post can contain up to 3,000 characters, and that a longer piece belongs in an article. It does not offer bold, italic, headings, or a real list. Line breaks are the formatting the composer actually stores. Everything else is characters you paste. Boykli writes and publishes that plain draft. It does not run a formatter, it does not insert bold letters, and it does not draw a "see more" preview. If you paste styled letters into the draft, they publish, because they are text. The app will not add them.
Logged-in visitors can browse Radar and Algorithm without a plan. Writing, images, carousels, Voice Agent, publishing, and scheduling need a paid seat. BYOK is $29 a month. Managed is $88 a month. Prices are the public ones on boykli.com.
Three layouts, and what a reader actually gets
| One dense paragraph | A full Unicode costume | A readable layout | |
|---|---|---|---|
| First line | Buried in a block the fold cuts at random | Often styled, so a screen reader spells letter names before the point | One plain sentence, short enough to finish before the fold |
| Breaks | Few or none | Many, sometimes held open with a hidden character | One blank line between blocks you typed yourself |
| Symbols | None | Bold, italic, monospace, script, arrows, and emoji on the same lines | One list marker. One bold label, if any. Sentences stay plain |
| Search, mentions, links | Work, if the words are Basic Latin | Break wherever a styled letter replaced a real one | Work, because the hook, the URL, the @name, and any hashtag stay plain |
| Editing later | Easy | Miserable. You cannot search for the word you wrote | Easy in the plain sentences. Only the label is awkward |
| What Boykli does with it | Can write this. You still have to break the lines | Will not apply it. Paste it yourself if you insist | Writes the draft. You add the breaks, then publish or queue |
The costume column is what most free "LinkedIn text formatter" pages sell, because the bold preview is the screenshot. The readable column is what still looks like a person wrote it on a train. AuthoredUp is the product that previews the fold and applies the letter swap inside an editor. That comparison is AuthoredUp alternative. Boykli is the writing and queue step before that pass, or instead of it when the draft does not need styled letters at all.
Write the draft before you style it
Create posts on a paid seat. Break the lines in the draft, then style a label if you still need one. BYOK is $29/month. Managed is $88/month.
boykli.com →Where the fold sits, and why a blank line moves it
LinkedIn does not publish the "see more" cut. AuthoredUp's measurement, which the formatting page also cites, puts it near 210 characters on desktop and near 140 on mobile. Treat both numbers as someone else's measurement, not as a LinkedIn specification. The range moves with the client, the font, and whether a blank line lands inside the opening.
A blank line in the first lines ends the preview early. That is useful when the first line is a complete sentence and the rest is support. It is a waste when the verb is on line two and the preview stops on a fragment. Keep the opening as one plain line, then the blank line, then the body.
Line breaks count toward the 3,000. So do spaces. A post that looks short because every sentence sits on its own line is burning characters on structure. That is usually the right trade. Burning them on a styled copy of the same sentence is not. Mathematical letters sit outside the Basic Multilingual Plane, so a counter that uses UTF-16 charges two code units per styled letter. LinkedIn's help page states the cap and does not state the unit. Leave a margin. The same warning, with the exact blocks, is on the formatting page.
Boykli's length labels are the writing target, not a second cap. Short is labeled 45–400 characters, medium 400–1400, long 1400–3000. The LinkedIn ceiling on the way out is still 3,000. A document post (a carousel PDF) has the same ceiling on its caption.
Characters that build a layout, and characters that only decorate it
| Character | Code point | What to use it for | What it costs |
|---|---|---|---|
| Line break | The Enter you press | The real structure. One idea, then a break | One character. A blank line is two. An early blank line shortens the fold |
| Ordinary space | U+0020 | Separating words | A run of spaces is not a column. The composer is not a monospace grid |
| No-break space | U+00A0 | Almost nothing in a post | Glues words so the line will not wrap. Some paste paths insert it without showing you |
| Braille blank | U+2800 | Some public formatters park it on an "empty" line so a gap survives paste | It is a character. A screen reader can announce it. It counts. It copies into a quote of your post |
| Zero-width space | U+200B | Almost nothing | An invisible hole in a word or a URL. You will not see why the link failed |
| Bullet | U+2022 • | A real list of three to five items | Not a list object. LinkedIn will not renumber or indent. Do not mix • with a second bullet style |
| Arrow | U+2192 → | One sequence, such as before and then after | A prefix on every line is a template, not a layout |
| Number plus period | "1." in plain digits | A sequence the reader must follow in order | You type the numbers. Nothing updates them when you delete step two |
| Sans-serif bold label | Mathematical bold, mapped on the formatting page | One short label at the start of a block | Two UTF-16 units per letter if the counter works that way. Screen readers often spell the mathematical name |
| Fullwidth Latin | U+FF21 and the letters after it | Nothing you want found or heard | Looks like spaced type. It is not the word you typed. Search, hashtags, and mentions miss it |
| Emoji | Various | One, when the sentence needs it | Many are two code units, and a skin-tone modifier adds more. Boykli's writer is told to skip them |
Script, fraktur, and double-struck letters sit in the same mathematical block as bold. They fail on a phone. Underline and strike are combining marks (U+0332 and U+0336) that collide or vanish on some clients. Box-drawing bars look sharp in a desktop screenshot and break into missing glyphs on a phone that does not have the font. A row of plain hyphens is boring and it survives.
If the list is the point of the post, three bullets in plain sentences beat a bold bullet, an italic bullet, and a monospace bullet. The hook is a sentence. Style it only after you would post that sentence with nothing after it.
Invisible characters a formatter leaves behind
Paste a formatted post into an editor that can show special characters, or into a small checker you trust, before it goes on a queue. Four leftovers show up constantly.
- No-break spaces from a word processor. Word and Google Docs use paragraph spacing for the visual gap and sometimes U+00A0 where you thought you typed a normal space. LinkedIn then either drops the gap or refuses to wrap a line. Retype the breaks in the Boykli draft or in any plain box. Do not paste straight from Docs into the formatter and then into LinkedIn.
- Braille blanks used as fake empty lines. U+2800 has no dots, so it looks empty in a proportional font. It is still U+2800. If your gap only exists because that character is on the line, delete it and press Enter twice. A gap you can delete by selecting "nothing" is a character.
- Zero-width spaces inside tokens. Formatters and some web editors insert U+200B so a long word will wrap. Inside a URL, a hashtag, or an @mention, that character is a break in the token. The composer will not link it. Strip zero-width characters from any line that must resolve.
- A styled letter inside a plain word. One bold letter left in a URL, or a fullwidth digit in a year, is enough to break the link or to make the year unsearchable. After you convert a label, read the characters on either side. The label should be a whole word. The sentence around it should be Basic Latin.
A practical test takes ten seconds. Copy the draft into a new note. Search for the word you think you wrote, in plain characters. If the search misses the label, that is expected. If it misses the hook, the URL, or the name you @mentioned, the formatter touched something it should have left alone.
A layout you can repeat without looking like a template
This is a skeleton, not a post. The words have to come from a specific situation. The shape is what keeps a formatter from taking over.
First line is one plain sentence with the point in it.
One short paragraph. Two sentences at most. No styled letters in this block.
Label in sans-serif bold, then a plain clause on the same line.
• First item, plain, one line if you can.
• Second item. Same marker. No indent.
• Third item. Stop.
One closing sentence. A plain URL if you need one. No styled hashtag.
What that skeleton refuses:
- A styled first line. The fold is the only part many people read, and a screen reader should get that sentence as a sentence.
- A blank line between the hook and a second half of the same sentence.
- Nested bullets. Leading spaces are not a stable indent.
- More than one bold label. If every block has a label, nothing has a label.
- A divider row of box-drawing characters between sections. A blank line is the divider.
- Hashtags, mentions, and URLs in anything other than plain characters. Boykli's writer is told to skip hashtags and emoji. That rule is part of the post generator. If you add a hashtag, you add it yourself, in plain characters, after the layout is done.
The same skeleton works under a carousel. The caption is still a post. The slides can carry the diagram, the table, and the code. How those slides are built is the carousel generator. Do not also Unicode-style the caption into a second copy of the slides.
A workflow that formats the layout, then three words
Do the writing in plain characters. Lay the lines out while the text is still ordinary. Run a text formatter only on the spans you marked. Styled letters are a bad editing surface: search misses them, and a model asked to "make this more readable" will often emit Markdown asterisks, which LinkedIn shows as asterisks.
- Write on Create posts. Pick the length that matches the idea. Short, medium, and long take a core idea of at least 10 characters. YouTube, article, series, and document take a source. The tone list depends on the mode. The draft comes back without a Unicode pass and without a layout pass. Why the draft starts from your material is the founder account.
- Read it as one column, the width of a phone. If a paragraph wraps to six lines, break it. If two sentences make one point, keep them together. One idea per block is the rule. One sentence per line for the entire post is how a draft starts to look generated.
- Put the hook on its own line. Then one blank line. Then the body. If the first line is not a sentence you would post alone, rewrite the line. The hook generator page is about that sentence. This page will not rescue a weak one with an arrow prefix.
- Add a list only where the order or the set is the point. Use • or "1." you typed. Three to five items. Leave the explanation of each item in plain words on that same line, or on the next line without a new marker.
- Mark at most one label per post, with a convention you will delete, such as square brackets. Convert only that span to sans-serif bold. Confirm the word "the" still has an h if you used italic anywhere. The mapping, including the Planck-constant h at U+210E, is on the formatting page. Do not convert the hook.
- Strip the leftovers. No U+00A0, no U+200B, no U+2800, no fullwidth letters, no styled character inside a URL or a mention. Paste into a plain note and search.
- Count. Include the line breaks. If the plain draft was already near 3,000, do not style a paragraph. Long mode is labeled up to 3,000 because that is the LinkedIn ceiling. Styling a block can push a UTF-16 counter over the cap while a naive letter count still looks safe.
- Put the finished text back into the Boykli draft if Boykli is going to publish it. The queue sends the text it has. Formatting a copy in another window and then publishing the older plain version is a wasted pass. Publish replaces the internal
=== CTA ===marker with a blank line and sends the rest as the LinkedIn commentary. Scheduling for the personal profile is covered in LinkedIn scheduling tool. Company pages are not a destination from Boykli.
Two refinements are available on each draft. Use them to cut a block or to force a line break in the writing, not to ask for bold. Managed accounts also have a monthly cap of 130 post refinements. BYOK has no Boykli generation cap. You still pay OpenRouter for chat and images, and Google AI Studio for Voice Agent.
When the layout should leave the post
Some structures do not belong in a 3,000-character plain field, even with a careful formatter.
- A table. Unicode box drawing will not hold columns once the phone wraps. A comparison belongs on a carousel slide or in an article.
- More than a single line of code. Monospace letters do not make a code block. Spaces stay ordinary, operators have no monospace form, and a reader cannot paste the line into a file and compile it. One token in a sentence is the limit. A real snippet goes on a slide or behind a plain URL.
- A profile headline. The 220-character line under your name is plain text in Edit intro. Unicode bold there reads as broken type in search. That line is a different job, covered on the headline page.
- An article, when you need real bold, real italic, and headings. LinkedIn's posting help points you to articles once a post exceeds 3,000 characters. The article toolbar is actual styling. The feed post is not.
If the draft is hard to read because the idea is thin, another write from a better source will do more than any symbol. The formatter is the last five minutes, and only on the words that need weight.
FAQ
What does a LinkedIn text formatter actually change?
It replaces characters. Bold, italic, and monospace in the feed are other letters that look like those styles. The composer does not gain a bold button. Layout still comes from the line breaks and the plain symbols you leave in the text. The code points for each style are on the formatting page.
Does Boykli format posts with Unicode?
No. Create posts returns a plain draft. Publish sends that draft, after removing the internal === CTA === marker. Paste styled letters into the draft yourself if you want them in the queued post. There is no fold preview. People who want the formatter and the preview in one editor are looking at a different product. AuthoredUp is the comparison for that job.
Why did my spacing disappear after I pasted?
Word processors store the gap as paragraph spacing, not as an empty line. LinkedIn keeps the Enter characters you actually have. Retype the breaks in a plain draft. Do not "fix" a missing gap with a braille blank or a no-break space. Those characters survive in ways you will not want when someone copies the post.
Do line breaks and styled letters count toward 3,000?
Line breaks and spaces count. LinkedIn documents a 3,000-character post limit and does not document the unit. Mathematical bold, italic, and monospace letters are outside the Basic Multilingual Plane, so a UTF-16 count charges two code units each. Many emoji do the same. Keep a margin, and do not style a draft that is already close to the cap.
Will a formatted hashtag, mention, or link still work?
No. Hashtags and mentions resolve on plain characters. A URL with a styled letter, a fullwidth character, or a zero-width space in it is not the URL you copied. Leave the hook, the @name, and the link in Basic Latin.
How many bold words should a readable post have?
One label, or none. The first line stays plain. Lists use a single marker. Italic is for a title or a short quoted phrase, after you have checked the letter h. Monospace is for one token. A post that is bold from the first word to the last is harder to read than the plain draft and signals a template.
Can I indent or make columns with spaces?
No. A run of ordinary spaces is not a stable indent, and a no-break space can stop the line from wrapping. Keep every line flush. If the content is a table or aligned code, put it on a carousel slide or in an article, and leave the post as sentences.
Do I need a paid seat to lay out a post?
You do not need Boykli to press Enter or to paste Unicode into LinkedIn. You need a paid seat before Boykli will write, publish, or schedule. Browsing Radar and Algorithm does not require one. BYOK is $29 a month with your own keys. Managed is $88 a month with platform models and monthly caps.
See Boykli
Plain LinkedIn drafts from your own material, then a personal-profile queue. Add the line breaks yourself. Style a label only after the draft is one you would sign.
boykli.com →