Viral Clip Score — the anatomy of a clip that stands alone

Most clips fail before the first caption appears: they assume context the viewer never had. The question this tool answers is narrow and brutal — if this moment is torn out of the video and dropped into a stranger's feed, does it still work? You paste the clip's text (the transcript slice that would be spoken in the cut), and it scores four properties of standalone moments — hook, clarity, payoff, quotability — combines them with fixed weights into one number out of 100, and offers a rewritten opening line you can use or ignore. Everything runs in your browser; nothing you paste is stored or sent anywhere.

One expectation to set immediately: this is a checklist, not a prophecy. No score predicts views, because views depend on distribution, timing, thumbnail, audience, and luck — none of which live in the text. What the score can tell you is whether the clip has the structural properties that give it a chance.

The four axes, defined precisely

Hook — does the opening grab? The first line is scored on explicit signals: whether it starts with one of the trigger openings the checker looks for (why, how, what, stop, never, "the truth", "I stopped", "we need"), whether the text contains a question mark, whether it contains a number, and how long it runs. The base is 45 points; trigger openings add 20, a question adds 15, a digit adds 10, and length contributes up to 15 more at half a point per word. The reasoning is that openings posing a question, naming a number, or taking a prohibition stance give a cold viewer a reason to stay through second three.

Clarity — is one complete thought present? This axis rewards the middle and punishes both extremes. A text between 12 and 40 words earns 95 total; anything shorter or longer earns 70. The window exists because a thought under 12 words is usually a fragment of a larger idea, while one over 40 words is usually two ideas wearing one coat. Standalone clips live in between.

Payoff — does the moment land somewhere? A base of 60, plus 15 if the text runs past 120 characters (enough room for a setup and a landing), plus 10 if it contains quotation marks — a quoted line inside the text is usually the thing the clip exists to deliver.

Quotability — could a stranger repeat it? A base of 55, plus 20 if the text is 28 words or fewer, plus 10 for quotation marks. The logic: ideas that fit in one breath with a sharp edge are the ones people carry out of the video and into their own conversations.

The weighting, and why hook dominates

The overall score is a fixed weighted sum, with a floor that keeps an empty input from scoring zero:

overall = hook × 0.35 + clarity × 0.20 + payoff × 0.20 + quotability × 0.15 + 70 × 0.10

That last term is the floor: seven guaranteed points, which is why the lowest possible overall is 60 rather than 0 (an empty paste scores hook 45, clarity 70, payoff 60, quotability 75, plus the floor — landing at exactly 60). The floor keeps the output interpretable: a terrible paste reads as "needs work," not as a system error.

Hook carries 35% of the weighted total because short-form feeds decide in the first seconds. A clip with a weak opening and a strong ending never reaches its ending; a clip with a strong opening and a modest ending at least gets watched far enough to be shared. Clarity and payoff share 40% because a hook that opens a loop nothing closes feels like clickbait within ten seconds. Quotability takes 15% because repeatable lines amplify a clip after the watch — they are what captions, comments, and stitches are made of.

The weights are constants, not a black box. You can reproduce any overall score by hand from the four subscores, which is the point: a score you cannot audit is an opinion, and this tool refuses to be one.

Reading your number: the three bands

The interface maps the overall score onto three honest bands:

• 85 and up — strong standalone moment. The text grabs, completes itself, lands, and fits in a breath. The recommendation is to try it as a 30–45 second cut, then verify against the footage that delivery matches the text.
• 65 to 84 — workable. Something real is here, but one axis is dragging. The usual suspects: the hook opens without a trigger, or the payoff arrives but the thought runs long. Tighten before cutting.
• Below 65 — not ready to travel. The paste is a fragment, an intro, or a run-on. It may be a great part of the full video and a poor clip; the fix is either restructuring or returning to the transcript for a different moment.

Because the floor holds the bottom at 60, treat the 60–64 zone as "the tool sees almost nothing to score" rather than as a judgment on your content.

Worked example: scoring one candidate by hand

Note for viral-clip-score: This is a hypothetical example — an invented clip used to demonstrate the arithmetic, not a measured result from a real creator.

Take this fictional paste: "Why my best video got zero views — I spent 40 hours on a script, and the one I filmed in a parking lot outperformed it 20 to 1." It contains 29 whitespace-separated tokens and 127 characters.

1. Hook: starts with "why" (+20), contains digits (+10), no question mark, and its 29 tokens contribute half a point each, capped — 14.5 here: 45 + 20 + 10 + 14.5 = 89.5, rounded to 90.
2. Clarity: 29 tokens sits inside the 12–40 window → 95.
3. Payoff: 127 characters clears the 120-character gate (+15); no quotation marks → 75.
4. Quotability: 29 tokens misses the 28-or-fewer window by one, and there are no quotes → 55.
5. Overall: 90 × 0.35 + 95 × 0.20 + 75 × 0.20 + 55 × 0.15 + 7 = 31.5 + 19 + 15 + 8.25 + 7 = 81 (rounded from 80.75).

An 81 reads "workable" — and the audit shows exactly why: quotability is starving because the thought is one token over the repeating window. Now add a quoted landing line: "…outperformed it 20 to 1. 'Ship the true version,' my editor said." The quote marks earn +10 on payoff and +10 on quotability in one stroke; the same run of arithmetic now returns payoff 85, quotability 65, and an overall of 84. That is the intended use of the tool: not to accept a number, but to read the four bars and intervene where the weakest one points.

The hook rewrite — what it does and the ethics of it

Below the bars, the tool offers a rewritten opening. The rewrite is deliberately tiny: it takes only the first sentence of your paste, capitalizes and adjusts its first word, and appends a framing tail. The substance of the sentence is never touched, and the rest of the clip is expected to stay verbatim.

Two ethical lines come with that feature. First, a hook rewrite may change emphasis but never meaning: if the speaker did not say it, the hook cannot promise it. Rewriting "we tried three encoders" into "the encoder nobody told you about" is a lie with better pacing, and viewers punish lies with the one metric that matters — they leave. Second, the rewrite is a suggestion generator, not a fact source: it cannot know whether the appended framing is true for your video. Use it as a shape to react to, then write the actual opening in the speaker's own truthful words.

What depresses each subscore, and the fix for each

Hook stuck in the 40s. The opening is narrative ("So we were talking about…") instead of positional. Move the claim to the front. The trigger list is a hint, not a spell — any opening that poses a question, names a number, or takes a stance will outperform throat-clearing.

Clarity at 70 with short text. Under 12 words means a fragment. Add the sentence that completes the thought — usually the one right after the paste in the transcript.

Clarity at 70 with long text. Over 40 words means two thoughts. Cut at the hinge and score each half separately; one of them is usually the real clip.

Payoff in the 60s. The moment sets up but never lands within the paste. Either the landing sentence was left out (include it) or the moment genuinely has no landing (it is context, not a clip).

Quotability in the 50s. The idea is there but too long to repeat. Compress to one sentence under 28 words, or accept that this clip travels by watch-through rather than by quotation.

Where this tool fits in a clipping workflow

Use it on candidates, not on the whole video. The productive sequence: mine the transcript for candidate moments (the companion tool on this site returns a ranked shortlist), then bring each finalist here to test whether it survives alone, then cut the survivors with padding and verify by ear. Clips that score high but fail the ear test are rare; clips that score low almost never surprise you in the edit. The score is cheapest at exactly the stage where mistakes are cheapest — before any cutting starts.

FAQ

Does a high score mean the clip will go viral? No, and the tool says so in its own interface. It measures standalone structure — the necessary condition — while virality depends on distribution and audience factors no text score can see.

Why does an empty paste score 60 instead of 0? The formula includes a fixed floor term, so the lowest possible overall is 60. Read that zone as "nothing to evaluate yet," not as a verdict.

Can I paste a clip with timestamps or speaker labels? You can; they count as words, which can push long pastes out of the clarity window and slightly distort the hook length term. For the cleanest reading, paste the spoken text only.

Is my text uploaded? No. Scoring happens locally in the browser tab, and nothing is stored server-side. The page works with your network disconnected.

Why did the rewrite change only my first sentence? By design. The tool touches only the opening and expects the remainder to stay verbatim, because rewriting substance would put invented words into a real speaker's mouth.

Which matters more — hook or payoff? For getting watched, hook; for getting shared, payoff. The weights favor hook because an unwatched payoff does not exist, but a hook-only clip burns trust fast. The honest target is both.

The same paste scores differently after I edited one word. Is that a bug? No. Several terms are threshold-based — the 12-word and 28-word windows, the 120-character payoff gate — so small edits near a boundary can move a subscore by its full bonus. The audit arithmetic above lets you see exactly which threshold you crossed.

Why the weights are fixed and always will be

A natural question: why not learn the weights from data, or tune them per platform? Three reasons, all about trust.

First, an adjustable score cannot be audited. The moment weights become parameters, "overall 81" stops being a fact you can reproduce and becomes an opinion you must take on faith. This page exists to be checked, not believed.

Second, platform-tuned weights would decay. What TikTok's feed rewarded in 2024 is not what it rewards now, and a score that silently re-tunes itself would make yesterday's advice wrong without telling you. Fixed weights age honestly: when the arithmetic stops matching reality, the mismatch is visible and the page can be revised in the open.

Third, the checklist is about invariants, not trends. Whether a thought completes itself, whether a stranger can follow it without context, whether it lands, whether it can be repeated — these properties mattered before short-form feeds existed and will matter after the current algorithms are forgotten. The weights encode that belief: hook first, because feeds are brutal; substance next, because feeds are not the whole story.

Interpreting subscore disagreements

The four bars rarely move together, and the disagreements are the most informative output the tool produces.

High hook, low payoff (90 / 60s). The classic clickbait shape: an opening that writes checks the moment never cashes. Viewers feel this within seconds and punish it with swipe-aways that hurt the channel more than a never-posted clip would. Fix by extending the paste forward until the landing appears — or accept that the moment is a teaser for the long video, not a standalone clip.

High quotability, low hook (75+ / 40s). A buried gem: the repeatable line exists, but the clip opens on warm-up. This is the cleanest structural fix in clipping — move the quotable sentence to the front verbatim, then let the original opening play as context. The same words, reordered, often jump twenty overall points.

High clarity, low quotability. The thought is complete but unmemorable — fine as a supporting clip inside a series, weak as a first contact with a cold audience. Consider pairing it with a moment that has an edge.

Everything mid-range (70s across the board). Usually a sign the paste is a healthy piece of a larger arc rather than a standalone unit. The tool is telling you the moment participates in a story without being one.

Editing patterns that raise scores without lying

Scores respond to structure, so structural edits are the legitimate lever. Four patterns cover most cases:

• Cold-open restructure. Take the payoff sentence, place it first as a two-second opener, then cut to the setup. Nothing is invented; chronology is the only thing that moves. This is the single highest-yield edit for low-hook clips.
• Context prepend. When a moment scores well but refers to something unnamed ("and that is when it broke"), prepend one sentence that names it. The clarity window forgives up to 40 words, so one context sentence usually fits inside the band.
• Trim to the thought. Long pastes lose clarity and quotability simultaneously. Find the hinge between the two ideas in the paste, cut there, and score each half. One half routinely outperforms the whole.
• Quote extraction. If the speaker said one line worth repeating, make sure it appears verbatim, in quotation marks, inside the paste. The scoring bonus for quotes is a proxy for what human quoters do: they look for the line they can lift unchanged.

What is not a legitimate lever: inventing claims, adding fake numbers, or rewriting the speaker's substance into something punchier than the truth. The hook rewrite feature is capped at the opening for exactly this reason.

Edge cases and unusual inputs

Very long pastes. The hook's length contribution caps at 15 points, so extra length stops paying early, while clarity actively penalizes pastes beyond 40 words. A 300-word paste is the wrong input shape: this is a per-clip tool, and long text belongs in the transcript miner instead.

Numbers everywhere. Digits add at most ten hook points once — one number and fifty numbers earn the same bonus. Stuffing a paste with statistics is therefore self-limiting, which is convenient, because it is also annoying to watch.

Questions inside the paste. Any question mark earns the hook bonus, including rhetorical ones. Rhetorical questions are legitimate hooks; a paste that is nothing but questions will score well and deliver nothing, which the payoff axis eventually reflects.

Punctuation-only fragments. A paste of "…?!" scores the floor and nothing else. The floor exists precisely so these inputs produce a readable minimum instead of nonsense.

Non-English text. The trigger-word list is English, so openings in other languages miss the +20 bonus even when they function as hooks. The remaining axes (question marks, numbers, length windows, quotes) still apply. Read scores on non-English pastes as partial information.

FAQ

Can two people paste the same text and get different scores? No. The scoring is deterministic arithmetic with no randomness and no personalization; same text, same numbers, every time.

What is the highest overall score possible? 100 in theory — hook, clarity, payoff and quotability all maxed plus the floor. In practice a paste would need a trigger opening, a question, a number, 28+ words for hook length, a 12–40 word count, 120+ characters, and quotation marks simultaneously, which is a contrived shape rather than natural speech.

Should I score the clip before or after adding captions? Before. Captions are presentation; the score measures the spoken structure. A weak structure does not get rescued by text on screen, and captioning a strong structure is where caption effort pays.

My clip scored 88 but flopped. Is the tool wrong? The tool never claimed otherwise — it measures whether a clip can stand alone, which it could. Performance depends on everything outside the text. The score told you the clip was sound; the feed owed you nothing.

Why a checklist beats memory when choosing moments

Creators picking clips from memory are subject to two well-known distortions. The peak-end effect makes the loudest moment and the final moment feel like the best moments, regardless of whether either makes sense out of context. Recency makes last week's episode feel richer than the one from two months ago, which is why back catalogs go unclipped. A structural checklist attacks both: it evaluates each candidate on properties a stranger would experience, not on what the recorder felt while recording.

The tool is not smarter than you — it is less biased than you, in the specific sense that its arithmetic has no memory of how the session felt. That is the entire value of scoring at all. When the checklist and your instinct disagree, the productive move is to articulate why: often the instinct is noticing delivery the text cannot see (keep the clip, trust the ear), but just as often the instinct is attached to a moment that needs twenty minutes of context to land (drop the clip, trust the score). The disagreement is the information.

Pre-flight: between a good score and publish

A score of 85+ earns the right to be cut. The remaining checks happen against footage, not text:

1. Play the moment with sound. Confirm the delivery matches — emphasis, pacing, no cough over the punchline.
2. Cut with padding. Two to five seconds of breathing room before the opening line and after the landing; cold opens cut mid-breath feel like errors even when the words are right.
3. Watch the first three seconds as a stranger would. Muted, small screen, no goodwill. If the opening line is not legible as a reason to stay, re-cut or re-caption.
4. Caption the spoken text, not a paraphrase. Paraphrased captions drift from the audio and read as wrong even to viewers who cannot say why.
5. Check the frame edges. Faces and text near the bottom and right edges risk living under platform interface overlays; the safe-zone checker on this site maps those regions per platform.

None of these steps change the score. They exist because the score covers structure, and structure is the floor a clip needs before the other crafts matter.