How to Time an Instagram Reel Script Before You Hit Record
ToolUpa’s Reels Shot-Clock builds a timed script in the browser. Word budgets use a speaking-rate estimate of about 2.5 words per second. That is not an Instagram limit, not a view forecast, and not a trending-audio picker. Nothing is uploaded or published.
Most weak Reels fail before the topic is even wrong. They fail on time. The speaker greets the algorithm, restates the title, apologises for lighting, and arrives at the point after the viewer has already flicked. A Reel is a shot-clock. Fifteen, thirty, sixty or ninety seconds is not a vibe. It is a budget of beats, each with a spoken line, an on-screen overlay and a visual job.
Instagram Reels Shot-Clock forces that budget. You name a topic and an audience, write one promise and one proof, pick a duration and a hook pattern. The first three seconds are reserved for the hook. Later beats get a slot budget derived from duration times a speaking-rate estimate. If a line overruns its slot, the beat is marked. You cut words, you do not “talk faster and hope”.
The first three seconds are the product
Open on motion or on a sentence that interrupts a habit. “Stop opening a meal-prep Reel with a greeting” is a hook. “Hey guys, so today I wanted to chat” is a stall. Interrupt, number, mistake, contrast and question are patterns, not magic spells. They exist so you do not default to a wave at the lens.
Do not put a logo sting in those three seconds. Brand comes after the viewer has a reason to stay. On-screen text in the hook should be six words or fewer. A phone cannot read a paragraph while the cut is moving. The overlay is a label, not a subtitle track of the entire speech.
If you cannot state the promise without the rest of the script, the promise is too big for the duration. A 15-second clip gets a hook, one point, and a close. It does not get two proofs and an objection. Adding duration is how you earn extra beats, not how you excuse a ramble.
Beats by duration
| Length | Beats | What you are allowed |
|---|---|---|
| 15s | Hook, one point, CTA | One idea. No second example. |
| 30s | Hook, promise, proof, CTA | One artefact on camera. |
| 60s | Adds a second proof and a recap | Two examples, still one thesis. |
| 90s | Adds an objection beat | Answer one doubt. Do not start a podcast. |
A 90-second Reel is still not a YouTube tutorial. If you need chapters, tags and a description field, you are in a different format. See the YouTube SEO checklist for long video metadata. Reels do not get that furniture. They get a clock.
Speaking-rate is an estimate, not a law
Two and a half words per second is a planning number for clear speech. Fast talkers exceed it. Careful talkers undershoot it. The shot-clock uses it so you can see overflow before you record. After you generate, read the script with a phone timer. If you cannot finish a beat before its end mark, cut the sentence. Do not shrink the font on the overlay to cram more speech.
Instagram does not publish a “maximum words per Reel” that this tool can enforce. Anyone selling a guaranteed word count as a ranking factor is inventing a rule. ToolUpa labels the budget as a speaking-rate estimate in the output notes. Keep that label when you export TXT for an editor.
Worked 30-second example
Topic: Sunday prep for 12-hour shifts. Audience: night-shift nurses who cook once a week. Promise: three boxes that survive a night locker. Proof: oats, a grain bowl, and a wrap that does not leak. CTA: save this and steal the locker map. Hook style: interrupt.
Seconds 0–3: stop opening this topic with a greeting. Seconds 3–8: the promise, on screen in six words. Seconds 8–24: the three boxes as proof, one visual at a time. Seconds 24–30: the save instruction on an end card, not a pile of hashtags over a face. If the proof paragraph is longer than the slot budget, you do not “speed up”. You drop the least visual box and keep two.
The caption fold under the video is a separate job. The shot-clock emits a 125-character fold from hook plus CTA so the still-and-caption view still makes sense if autoplay is off. Write the full caption in Caption Fold Writer if you need keyword checks and hashtag placement. Do not treat the timed speech as a caption dump.
Visuals are beats, not decoration
Each slot in the tool has a visual line: hard cut, artefact, close-up, objection cutaway, stacked recap titles, end card. Those are assignments. If every beat is “talking head in the same kitchen”, you have not used the clock. B-roll is not stock footage of a smiling stranger. It is the locker, the wrap, the actual mess.
End cards should carry the next step in the same words as the spoken CTA. If you say “comment PREP” and the card says “link in bio”, you split the ask. One ask per clip. Follows, comments, saves and profile visits are different jobs. Pick one.
What this workflow will not do
It will not pick trending audio. Sounds live in Instagram and change by region and week. A static site that lists “this week’s audio” is either stale or scraping. ToolUpa times the spoken track you already plan to say. Add music in the app after the words fit.
It will not forecast views, watch-through or a “viral score”. Those numbers, if they exist, live in Insights after you publish. If you want arithmetic on Insights you already copied, use the Engagement Ledger. Do not paste a fantasy view count into a client deck and attribute it to a script generator.
It will not replace a storyboard for carousels. Carousels are still slides. If the idea is a sequence of stills, use Carousel Storyboard Maker. A Reel script with six overlay lines is not a carousel with six readable slides.
Recording against the clock
Print or keep the TXT export next to the lens. Record beat by beat if you cannot land a full take. The cut points are already in the timestamps. Do not record a three-minute ramble and “find the Reel in the edit”. That is how you keep the greeting you meant to kill.
When a beat is marked over slot, cut adjectives first, then clauses that repeat the promise. Keep proper nouns the audience already uses. “Night locker” stays. “Incredibly easy and amazing” goes. If the proof still overruns after that, the duration is wrong. Move to 60 seconds or drop a proof.
Watch the take on mute. If the overlays and visuals still carry the thesis, the clip can survive a muted autoplay. If mute makes it a mystery, you leaned on speech the overlay cannot hold. Rewrite the six-word labels before you shoot again.
Checklist
- Duration chosen before the first sentence.
- Hook in three seconds, no greeting, no logo sting.
- One promise. Extra duration buys extra proof, not extra theses.
- Spoken words checked against the speaking-rate budget, then timed out loud.
- Overlays at six words. Visuals assigned per beat.
- One CTA, matching the end card.
- No trending-audio claim and no view forecast.
Time is the scarce material. Topic, audience, promise and proof only matter if they fit the clock. Generate the lattice, cut the overruns, then pick up the phone.
FAQ
Can I use this for Stories?
Stories have a different clock and a different UI. This tool is for Reels durations. You can steal the hook discipline. Do not pretend a 90-second script is a Story.
Is 2.5 words per second too slow?
It is a planning average. Time yourself. If you speak faster, you get more words per slot, but overlays still cap at six words.
Will ToolUpa upload the video?
No. There is no video encoder and no Instagram login.