The spec for LinkedIn
LinkedIn treats video as a quick-scan, mobile-first format. Audiograms can win when they respect the platform's constraints and habits.
- Aspect ratio - 1:1 square or 9:16 vertical perform best in feed. Use 1080x1080 or 1080x1920.
- Frame rate - 24 to 30 fps is smooth and efficient.
- Encoding - MP4 with H.264 video and AAC audio remains the safest choice.
- Duration cap - LinkedIn supports lengthy uploads, but attention decays fast. Aim for 45 to 60 seconds for most audiograms, up to 90 seconds if the content is undeniably strong.
- Sound-on vs sound-off - Auto-play is muted. Assume viewers start silent. Your first 3 seconds must read visually without audio.
- Captions - Always on. High contrast. Two lines max per card. Burned-in captions are the most reliable across devices. You can also attach an SRT, but do not depend on it for the hook.
- Thumbnails - A bold title card or the first frame matters because LinkedIn often shows the opening frame as the visual in feed.
The structure that works
Ship a tight 55 second audiogram that hits every beat. This structure has shipped successfully across dozens of LinkedIn posts in 2026.
0:00 - 0:03 Hook frame
- Big headline, 1-2 lines, no jargon. Example: "We cut our cloud bill 27 percent with one setting."
- Waveform visible immediately. Subtle brand color bars.
- Speaker name or role in a corner if relevant.
0:03 - 0:12 Context in one breath
- Why this matters to LinkedIn readers - time, money, team velocity, quality.
- One statistic or constraint. Example: "Most teams don't audit egress fees until Q3."
0:12 - 0:40 Value payload
- Micro-beat A - the principle. 10 to 12 seconds. Explain the lever you pull. Example: "Disable cross-zone replication for non-critical workloads."
- Micro-beat B - the step-by-step. 12 to 16 seconds. Three steps max. Clear verbs: "Audit, toggle, monitor."
- Micro-beat C - one pragmatic caveat. 6 to 8 seconds. Example: "Keep it on for stateful services."
0:40 - 0:52 CTA that respects the feed
- Offer a logical next step. "Comment 'Egress' and I'll DM our checklist." or "Link in comments."
- Show logo small. Include a discreet URL only if it is short and readable.
0:52 - 0:55 End frame
- Hold for 2 to 3 seconds. High contrast background, brand lockup, and a short reminder: "More engineering ops tips weekly."
- Do not fade to black instantly. Give the algorithm a frame that looks good in preview.
Hooks that earn attention
Hooks are compact, visually scannable lines that make sense without audio. Use these formulas and test variations.
Formula 1 - The measurable win
Template: "We cut [metric] by [number] with [one change]."
Examples:
- "We cut deploy rollback time by 68 percent with one flag."
- "We cut cloud egress fees 27 percent by moving one bucket."
Formula 2 - The hard truth
Template: "If you're doing [practice], you're wasting [resource]."
Examples:
- "If you're autoscaling on CPU alone, you're wasting money."
- "If you're ignoring cold starts, you're wasting user trust."
Formula 3 - The 3-step move
Template: "Fix [pain] in 3 steps: [step 1], [step 2], [step 3]."
Examples:
- "Fix noisy alerts in 3 steps: group, throttle, escalate."
- "Fix slow pull requests in 3 steps: label, auto-assign, gate."
Formula 4 - The myth vs reality
Template: "Myth: [belief]. Reality: [concise truth]."
Examples:
- "Myth: bigger instances mean faster builds. Reality: fix I/O limits."
- "Myth: more dashboards mean more insight. Reality: fewer, clearer SLOs."
Formula 5 - The checklist callout
Template: "Before [action], run this checklist."
Examples:
- "Before migrating regions, run this checklist."
- "Before enabling autoscaling, run this checklist."
Brand + voice
Audiograms do not win on novelty alone. Viewers remember consistent colors, type, layout, and a point of view. A strong brand kit plus a clear voice beats any single viral post because it compounds across your feed. The repetition builds recall, which drives engagement and inbound.
Define your kit:
- Primary and secondary colors - map them to different elements. Primary for headlines, secondary for waveform bars, neutral dark for caption background.
- Type system - one display weight for hook headers, one readable weight for captions. Keep kerning and line-height consistent across videos.
- Logo lockup - fixed position, 80 to 100 px width in 1080x1080. Do not animate it every time. Make it a reliable anchor.
- Motion vocabulary - one or two transitions only. Hard cut or a 120 ms slide-up. LinkedIn is not a cinema.
Tools like HyperVids make this reliable by letting you create a per-project brand kit once, then stamping it onto every audiogram. Colors, fonts, caption box style, waveform type, logo position, and CTA frames remain consistent while your content changes.
Voice matters equally. Choose your stance: "pragmatic engineer," "data-in-hand operator," or "always-testing marketer." Put a sharp POV into every hook and keep jargon low. Your viewer should be able to repeat your point in one sentence.
Captions + accessibility
Assume sound-off by default. Captions carry your message. Make them readable, accessible, and safe in the LinkedIn feed.
- Always-on captions - burn them in for the hook and the core beats.
- Contrast - WCAG AA 4.5:1 minimum. Practically, white text on charcoal or brand-dark with 80 to 85 percent opacity works well.
- Characters per line - cap at 32 to 38 characters. Two lines max. Shorten nested clauses. Replace commas with periods when it helps rhythm.
- Font sizing - in a 1080x1080 square, 52 to 62 px for captions is readable. In 1080x1920 vertical, 56 to 68 px. Test on a phone at arm's length.
- Safe margins - keep 80 to 100 px horizontal padding and 120 px above/below captions. LinkedIn UI can overlap the bottom edge on some devices.
- Timing - 1.5 to 2.5 seconds per caption card. Do not time to syllables. Time to ideas.
- Punctuation - period and emulation of natural breaks. Avoid ellipses and emojis in captions. Emojis can reduce clarity for screen readers.
- Screen reader support - if you attach SRT, write full sentences. Use speaker labeling for multi-voice edits. Example: "Sam: We toggled this flag and cut requeues."
A sample HyperVids prompt
Below is a realistic prompt that sets your brand context and asks for a LinkedIn audiogram. Keep it short, concrete, and testable.
/hyperframes audiogram linkedin
Brand:
Name: Acme Cloud
Palette: #2D71F7 primary, #0B1220 dark, #F2F6FF light
Fonts: Inter Semibold for headers, Inter Regular for captions
Logo: acmecloud.svg top-left, 92px width on 1080x1080
Waveform: vertical bars, brand blue, subtle glow
Caption box: rounded, dark #0B1220 at 85% opacity, 2 lines max, 34 chars per line
Video:
Ratio: 1:1 at 1080x1080, 55s total, 30fps
Motion: slide-up 120ms on captions, hard cuts between beats
Content:
Hook: "We cut cloud egress fees 27% with one setting."
Context: "Teams ignore egress until Q3, then it hurts."
Value:
- "Principle: stop cross-zone replication for non-critical workloads."
- "Steps: audit buckets, toggle per policy, monitor exceptions."
- "Caveat: keep replication for stateful services."
CTA: "Comment 'Egress' for the checklist - link in comments."
Audio:
Track: founder.mp3, clean up background, normalize to -14 LUFS
Output:
Render MP4, burned-in captions, first frame uses hook as thumbnail
When you trigger this through HyperVids, the /hyperframes skill will apply your brand kit, lay out the beats to fit LinkedIn's time box, generate captions that honor line-length rules, and render a square audiogram. With a connected Claude CLI subscription, you also get crisp copy alternates for the hook and CTA in one pass.
Common failure modes
- Muted hook - opening visuals require audio to make sense. Fix it by adding a bold headline that stands alone.
- Too much text - captions that run three or more lines per card feel crowded. Tighten to two lines max and split across more cards.
- Low contrast - gray text on light brand backgrounds is unreadable on mobile. Increase opacity or shift to a darker overlay.
- Jargon-heavy voice - "cost optimization via multi-az replication policy" is not scannable. Rewrite to "stop cross-zone copies for non-critical data."
- Long runtime - 2 minute audiograms underperform in feed. Trim to one idea and ship at 45 to 60 seconds.
- Weak CTA - "Learn more" is vague. Offer a specific next step that fits LinkedIn, like a checklist or a short doc in comments.
- Inconsistent brand - shifting colors and fonts post to post lowers recall. Lock your kit and reuse it.
- No testing - posting once and hoping. Batch three variants on hooks, then measure saves, comments, and completion rate.
Conclusion
LinkedIn rewards clarity, brevity, and consistency. A tight audiogram that hooks visually, delivers one practical insight, and offers a concrete next step will outperform generic talking-head clips. Build a reusable brand kit, ship multiple versions, and optimize your captions for sound-off viewing. Do this weekly and your feed becomes a reliable acquisition surface.
If you prefer a faster path from idea to output, HyperVids makes the workflow repeatable for teams that care about brand and speed. The difference is not magic - it is a clean spec, a dependable structure, and deliberate iteration.
FAQ
Should I post square or vertical on LinkedIn for audiograms?
Square 1:1 is the safest default across desktop and mobile. If your audience skews mobile, test 9:16 vertical. Keep captions inside safe margins so the UI never overlaps text.
Do I need an SRT file if I burn in captions?
Burned-in captions guarantee the hook works in feed. An SRT still helps accessibility and search. Attach it when possible, but never rely on it for the opening message.
How many times can I reuse the same hook?
Twice with variations is fine. Change the stat, narrow the audience, or shift the outcome. Tools like HyperVids can generate alternate hooks quickly so you keep testing without losing your brand consistency.