YouTube Title Counter

The cap, as of mid-2026

A YouTube title accepts 100 characters. That figure is a product decision, not a standard: it reflects YouTube's published rule as of mid-2026, and when the platform changes it, this page's copy and the preset above are updated together. Everything else on this page follows from one distinction. The cap is a field size — the most the title can hold. How much of a title a viewer sees on a particular device is a display question, and the two are not the same number. The counter above measures the first exactly and makes no claim about the second.

Read next: Truncation and folds: what the checker cannot see — why YouTube's device-dependent truncation, Instagram's feed fold and Google's snippet cut are described on this site rather than simulated, and how to draft around each. More in the guides.

What is counted

The unit on this page is the Unicode code point — the unit most platforms document their caps in, and the one the methodology page sets out in full. In practice: an accented letter such as é is one, a CJK character is one, and an emoji such as 🎉 measures 1 in a limit check, although JavaScript stores it in 2 UTF-16 units. Spaces, punctuation, dashes and line breaks are code points too, and each counts as one — there is no category of character the meter skips. The Characters stat card shows the same code-point total the remaining figure is derived from, so the two never disagree. The Words card is a secondary count using Unicode-aware word detection, in which internal hyphens and apostrophes keep a word whole; "Torch-Free Crème Brûlée: A No-Fuss Method" is 6 words by that rule and 41 characters by the one that matters here.

The one wrinkle is the multi-part emoji. What looks like a single family glyph, 👨‍👩‍👧, is several code points held together by invisible joiner characters, and the counter reports what is there rather than what is drawn: 🎉 on its own measures 1, while 👨‍👩‍👧 measures 5. Platforms do not agree with one another on how to count such sequences, which is why a borderline title that contains one belongs in YouTube's own composer for its final check — a point this page returns to below.

Worked examples, computed by the engine

Every number in this section comes from the same function that runs under the meter above, called on the sample text at build time. Start with a title that fits: "Soldering for Beginners: Clean Joints on the First Try" measures 54 code points, leaving 46 of the 100 unused. Now the kind of draft most titles start as — "The Complete Beginner's Guide to Soldering Clean Joints Every Time: Tools, Technique and the Mistakes to Avoid" — which measures 110: 10 over the cap, so the readout switches from "remaining" to "over" and the bar fills. A title that overruns the field has to lose characters before anything else is worth deciding. Cutting the preamble rather than the content words gives "Soldering Clean Joints Every Time: Tools, Technique, Mistakes to Avoid", which measures 70 and leaves 30 to spare. That is the whole edit: the subject leads, the detail follows, and the draft is inside the field with room.

Accents and scripts do not change the arithmetic. "Crème brûlée at home: the torch-free method" measures 43, each precomposed accented letter counting as one. A title written in Japanese — "はんだ付け入門：きれいな接合部を作る方法" — measures 20, one per character, so a CJK title spends the cap at the same rate as a Latin one under a code-point rule. Emoji follow the same rule: "Weekend Sourdough 🍞 Start to Finish" measures 35, and swapping the emoji for an em dash — "Weekend Sourdough — Start to Finish" — measures 35. The emoji and the dash cost the same — 🍞 alone measures 1, the dash alone 1 — even though the emoji takes 2 UTF-16 units to store and the dash 1.

Two drafts that look identical can differ by one. "Soldering Basics" measures 16; the same title with a line break after it measures 17, and with a trailing space 17. Neither is visible in the box, and both count. When a count reads one higher than a by-hand tally, the invisible character at the end is the first thing to check.

Device-dependent truncation: described, not modeled

This site sorts platform behavior into two kinds: rules its engine runs, and rules it only states. A rule goes into the second group when running it would mean either guessing at what happens on the platform's side, out of view, or asserting a figure the platform can change without notice; the modeled-versus-described guide sets out the reasoning and the methodology page the full list. The code-point count against the published cap is in the first group — it is what the meter does. YouTube's device-dependent title truncation is in the second, described but not modeled: the same title is shortened at different points on different devices and surfaces, and a page that claimed to know the exact cutoff for your viewer would be inventing a number. So this page states the quirk and tells you which direction it skews the count you see here.

The direction is one-way. Truncation can only make the visible part of a title shorter than what the field holds; it never makes a title that exceeds the cap fit. A draft that passes here has therefore cleared the field and nothing more — some device may still show less of it. A draft that fails here exceeds the cap regardless of device. Read the meter as an upper bound on what any viewer sees, never as a promise that all of it is shown. Because this page cannot say where a particular surface will cut, the practical response is a drafting habit, not a second meter: let the words that carry the meaning come first, so that however much of the title a surface shows, the part it shows makes sense on its own. The truncation and folds guide works through this for YouTube alongside the other two display cuts this site describes.

Where the checker stops

The meter answers one question: how many code points are in the box, and how that compares with 100. It does not know which device your viewer holds, so it does not predict where a title is cut. It counts whatever you type; the composer is the final authority on anything beyond the count. It does not apply the weighted counting described on the X counter, which is specific to that platform; here the published cap is checked as a plain code-point count. And it is not a server: counting runs as JavaScript in your browser, there is no counting endpoint, and you can confirm in your browser's network tab that typing generates no requests. The engine's own ceiling is two million UTF-16 units — far beyond anything a title field will see.

How to read a borderline draft

A draft within a few characters of the cap deserves three checks before you trust it. First, look for joined emoji sequences such as a family emoji, because the counter reports their code points and platforms disagree on how to count them; a title whose margin depends on one of these should be confirmed in YouTube's composer. Second, look for invisible characters at the end: a line break or trailing space pasted in with the title costs one each (the text box delivers a line break as a single line feed), as the examples above show, and is one reason a by-hand tally and the meter can differ by exactly one. Third, remember what the cap is: clearing it means the title fits the field, and the device-dependent truncation described above still applies to the visible part. When the readout reads "over", the only remedy is to remove characters; the meter does not trim anything for you, and it should not, because which words to lose is the decision that matters. Until you have verified a consequential title against YouTube itself, treat this counter, like any third-party counter, as a drafting aid rather than a guarantee.

Common misreadings

The first misreading is treating 100 as what viewers see. It is what the field holds; what is shown depends on the device, as the previous sections explain. The second is importing the SMS trap onto a code-point cap. On the SMS checker, a curly apostrophe pasted from a word processor is the classic silent flip: it sits outside the GSM-7 alphabet, so the whole message switches encoding and budget. Here there is no such flip. "Don't Buy a Soldering Iron Until You Watch This" measures 47 with a straight apostrophe and "Don’t Buy a Soldering Iron Until You Watch This" measures 47 with a curly one, because each apostrophe is a single code point. The same two strings handed to the SMS engine read GSM-7 with a budget of 160 and UCS-2 with a budget of 70 — the flip is real, but it belongs to SMS, not to a title cap. The third misreading is expecting an emoji to cost two: that is its UTF-16 length, not its code-point count, and this counter reports code points; 🎉 is 1 here. The fourth is reading the Words card as a limit. It is a secondary count for orientation; the cap applies to characters only. And the last is assuming the cap is permanent. It is a published rule as of mid-2026, and the page says so precisely because such rules change.

Frequently asked questions

Counting is in Unicode code points; where YouTube's surfaces truncate varies by device. Nothing you type is transmitted or stored. See the methodology page.