Guide
What should the cover say? Ten short lines, drawn on the frame
A cover frame gets read at the size of a fingernail, in a grid of other videos, by somebody who has already decided to scroll. That leaves room for two to five words. Type what your video is about and the planner above writes ten of them, from eight patterns that work on a cover, and draws each one at real 9:16 proportions in your browser.
Cover text planner
Runs in your browser · no account needed
Ten cover lines, drawn on the frame they have to fit
Say what the video is about and you get ten short lines — two to five words each — pulled from eight patterns that work on a cover frame. Every one is drawn at real 9:16 proportions, at the size and position you pick, with the strip the feed covers marked.
Text size
reads in the feed grid
Position
across the face
Example · curiosity gap
“What nobody tells you” — one of the ten. Type a subject above and the other nine name it: the mistake, the number, the question, the warning.
The shaded strip is roughly where TikTok, Reels and Shorts draw their own interface. The ghost behind the words is where your face or your product goes.
This designs the words, not the picture. A cover frame still needs a face looking at the lens or one clear object filling most of it — text over a busy, unreadable frame is still unreadable. Lines in [brackets] are yours to fill: put your own street, town or venue in them.
Keep the words above the shaded strip, and check the frame you actually pick against the real margins at the safe-zone checker. Nothing you type here is sent or stored anywhere — close the tab and it is gone.
- Cover lines written for your subject
- 10
- Words that still read at thumbnail size
- 2-5
- Patterns the lines are drawn from
- 8
The cover is an advert for the video
On a profile grid, a search result or a suggested row, your video is a still frame the size of a stamp with everything else on the screen competing for the same thumb. The cover is the only part of it doing any selling.
That is a different job from the video's opening line. The opening line has a second or two of attention already granted; the cover has to earn the tap before anybody has granted anything.
Which is why the word count is so brutal. Two to five words is not a style preference, it is what a person can take in from a frame at that size while moving. Anything longer is decoration that reads as a grey smudge.
The words also have to survive a crop and an overlay. Every feed app draws its own interface across the bottom of the frame — the account line, the caption, the buttons — so text parked there is text nobody sees.
The planner above puts those two constraints on screen instead of in a paragraph: every line it writes is inside the word budget, and every one is drawn on a 9:16 frame with the strip the interface covers shaded in.
What it will not do is invent a subject for you. Type three or four words about the actual video and the ten lines name it; type nothing and it says so rather than handing back ten lines about nothing.
Eight patterns that fit on a cover
There are not a hundred kinds of cover text. There are about eight, and every strong cover you have scrolled past is one of them wearing a different subject.
Knowing the eight matters more than knowing any single line, because it turns a blank frame into a choice: is this video a count, a warning, a mistake, or a before and after? The answer usually decides the words in one go.
The planner deals one line from each of the eight in the first ten, on purpose. Seeing your own subject in all eight registers at once is the fastest way to notice that the video you thought was a how-to is really a warning.
A pattern is not a guarantee. A number promises a finite watch and then has to deliver three actual things; a curiosity gap that withholds the topic itself is just vague, and vague loses to specific every time.
The rule underneath all eight: name the thing. A cover that says 'watch this' is asking for a favour. A cover that says '3 gutter mistakes' is offering something, and the offer is what gets tapped.
The eight cover-text patterns, and where each one belongs
One line from each of these lands in the first ten the planner writes for your subject.
| What it promises | Where it earns its place | |
|---|---|---|
| Number | A finite watch — three things, then done | Lists, checklists, anything with steps |
| Question | The thing they were already half-wondering | Comparisons, buying decisions, myths |
| Mistake | An error they might be making right now | Technique, DIY, money, anything with a wrong way |
| Before / after | A change, visible in the frame itself | Trades, transformations, renovations, training |
| Named place | That this is about here, not everywhere | Local services, property, events, food |
| Blunt claim | An opinion stated flat out | Reviews, hot takes, anything with received wisdom |
| Curiosity gap | That something is being held back | Stories, behind the scenes, industry secrets |
| Warning | A cost, delivered before they pay it | Contracts, purchases, first attempts |
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Two to five words, and where to put them
Size and position are the other half of the decision, and they are the half people guess at. The planner above gives you three sizes and two positions and draws the result, because the only honest way to judge legibility is to look at it small.
Big text is not automatically better. A line that fills the frame reads from across the room, but it also buries whatever is behind it — and on a talking-head video, what is behind it is your face, which is doing its own persuading.
Middle placement puts the words across the subject and reads as a statement. Top placement leaves the face clear and reads as a label. Neither is wrong; a video whose picture carries the story wants the label, a video whose words carry it wants the statement.
The bottom third is not a choice at all. That is where the feed draws its own interface, and the exact margins differ by app — which is why the shaded strip in the previews is a hint rather than a measurement. Check the frame you actually pick against the real per-app margins at /tools/safe-zone-checker.
One more thing the previews show without saying: white words need a dark edge. Text with an outline or a shadow survives a bright sky, a white wall and a pale shirt; text without one disappears into all three.
What a cover line cannot do
This designs the words. It does not design the picture, and the picture is most of the work.
A cover frame needs a face looking down the lens or one clear object filling most of the frame. Text over a busy, mid-motion still is unreadable no matter how good the line is, and the best line in the world cannot rescue a frame where nothing is legible.
So pick the frame first. Somebody's eyes open and pointed at the camera, a hand holding the thing the video is about, the finished result — then put four words on it.
The type here is your device's, as well. Every platform's cover editor draws text with its own fonts, weights and outlines, so read the previews for length, size and placement rather than for exact pixels.
And a cover is a promise the video has to keep. Writing a line that overstates what is in the first ten seconds buys one tap and costs the thing that actually compounds, which is people believing your covers.
The planner leaves brackets in one place on purpose: it does not know your town. When a line comes back with a bracketed word, that word is yours to fill in.
From a cover line to the video under it
The cover is the last thing you choose and the video is the thing it has to be true about. Cutroom makes the second part.
Film one take of up to three minutes on your phone, upload it, and it comes back as a finished 9:16 video: cut from the transcript, silences trimmed, captions on the words as you actually said them, b-roll over the phrases you highlight.
The line you just picked usually has a second job. Writing the opening headline — or picking one the machine drafted — is one of the things you can change in the editor, and the cover line and the opening headline saying the same thing is what makes a video feel deliberate.
Picking the cover frame itself happens where you post. TikTok, Instagram and YouTube each let you scrub to a frame and add your text when you upload, which is where these ten lines are meant to end up.
The pricing is one sum. A batch is 100 credits and the export is 20 credits per output minute, so a 60-second finished video is about 120 credits.
Three finished videos are free inside your first seven days, no card, on the 600 credits granted at signup, and they export with a Cutroom mark across the middle. Lite is $19.99 a month for 1,250 credits — about ten finished one-minute videos, more when they run shorter — and cancelling is one click.
Questions people ask
- What should I write on a video thumbnail?
- Two to five words that name the thing and make one promise: a count ("3 gutter mistakes"), a mistake, a question, a warning, a before and after, a place, a blunt opinion, or a stated gap. Type your subject into the planner above and it writes ten of them, one from each pattern, so you can see which register your video is actually in.
- How many words should cover text be?
- Two to five. A cover frame is read at the size of a stamp in a grid of competing videos, so the ceiling is what somebody can absorb at a glance while scrolling. Everything the planner writes is inside that budget, and it shows each line on a 9:16 frame so you can check it at real proportions.
- Where should the text go on a Reels or TikTok cover?
- Across the middle or near the top, never in the bottom third. Every feed app draws its own interface over the bottom of the frame — the account line, the caption, the action buttons — so text there gets covered. The previews above shade that strip; the exact per-app margins are at /tools/safe-zone-checker.
- Do Reels and TikTok covers need text at all?
- Not always. A cover with a strong face or one clear object can carry itself, and text over a busy frame makes things worse rather than better. Text earns its place when the picture alone cannot say what the video is about — which is most explainers, lists and before-and-afters, and few pure reaction clips.
- Is my subject uploaded when I use this tool?
- No. The patterns ship with the page and the lines are written in your browser, with no account needed. Nothing you type is sent or stored anywhere — close the tab and it is gone.
Pick the frame, pick four words, and make sure the video underneath keeps the promise they make.