Most launch videos fail for a boring reason. The team spends three weeks on a 90 second hero film, ships it on launch day, and then discovers they also needed a 15 second teaser for X, a vertical cut for Instagram, and a silent autoplay loop for the top of the landing page. None of those exist, so the hero film gets cropped badly and posted everywhere.

A product launch video is not one asset. Planning for that from the start changes what you should build and what you should build it with.

## The three cuts you actually need

For a software launch, the useful set is small and specific.

**The hero page loop** sits above or beside your headline. It runs 8 to 15 seconds, loops without a visible cut, and plays muted with no controls. Nobody watches it deliberately. It exists to show that the product is real and to give the eye something to settle on while the reader processes your headline. Text inside it has to be legible at a glance because it may only get one pass.

**The teaser** is 15 to 30 seconds for social. It has one job: make someone stop scrolling and click through. It should show the single most surprising thing your product does, not a tour. If your product turns a two hour task into two minutes, that comparison is the teaser.

**The feature reveal** runs 60 to 120 seconds and lives on your launch post, your Product Hunt page, and your docs. This is where you can actually explain something. It carries voiceover or captions and walks through what the product does in the order a new user would encounter it.

Everything else is a variation on these. A vertical cut of the teaser, a 6 second bumper, a GIF for your changelog. Once the three exist, the variations are crops and trims rather than new productions.

## Real product UI beats stock B-roll

The instinct when you cannot edit is to reach for stock footage: someone typing on a laptop in a bright office, a slow push in on a coffee cup. It fills time and it looks professional in the sense that it is well lit.

It also tells the viewer nothing, and audiences have gotten very good at recognising it. A launch video built from stock reads as a launch video for a product nobody wanted to show.

Your actual screen recording is more interesting than any stock clip, even if it is a bit rough. The fix for rough screen capture is not to replace it. The fix is to add graphics on top: a callout that points at the thing you want noticed, a number that counts up, a label that names the step. That combination reads as confident.

## What to animate versus what to record

Record the parts where the product does something. Animate the parts where you are making a claim about it.

| Content                                    | Approach                                        |
| ------------------------------------------ | ----------------------------------------------- |
| A user completing a task in your app       | Screen recording, sped up 1.5x to 2x            |
| "3x faster than the manual process"        | Animated comparison bar or a side by side timer |
| Where your customers are                   | Animated map with markers                       |
| A metric you are proud of                  | Animated stat callout with a count up           |
| How the pieces of your system connect      | Animated diagram, built up one node at a time   |
| What changed between the old way and yours | Animated before and after chart                 |
| Your roadmap or launch sequence            | Animated timeline                               |

A comparison claim built as a graphic instead of left as narration. Two quantities, sized against each other.

The screen recordings you can capture with QuickTime, OBS, or your operating system's built in recorder. The graphics are the part that historically required either a motion designer or 30 hours of After Effects tutorials, which is why most launch videos skip them and end up as a screen recording with a music bed.

[Moshion](/create) generates animated content from a text description: charts, maps, timelines, text animations, text highlights, and complex animated concepts. Export as MP4 and drop into any editor.

## A one day production plan

### 1\. Write the feature reveal script first

Write the 90 second version before anything else, even though you will publish the teaser first. The long version forces you to decide what the product actually is. The teaser is then a cut from it rather than a separate creative exercise.

Keep it to roughly 150 words per minute of video. Mark every sentence where you are stating a number, a comparison, or a relationship. Those marks are your graphics list.

### 2\. Capture the screen recordings

Record at the resolution you will publish at or higher, and record more than you need. Do each interaction as a clean single take rather than trying to stitch. Move the cursor deliberately and pause for a beat after each click so there is room to place a callout later.

One practical detail that saves a lot of pain: set your display to a clean state first. No notifications, no half filled test data, no email address that reveals your staging environment.

### 3\. Generate the graphics

Work from the marks you made in the script. Each one becomes a short animated clip, usually two to five seconds, that either sits over the screen recording or cuts in as a full frame.

This is where a text to animation tool earns its place. You describe the graphic you need, get an MP4 back, and drop it into your timeline. A stat callout that counts from zero to your number, a bar chart comparing your product to the manual process, a map showing your first hundred customers.

Export the ones that sit over screen recordings with a transparent background so you can composite them without a matte. Export the full frame ones as standard MP4.

A count up, the kind that sits under a claim your narration is making. Two to four seconds, then back to the product.

### 4\. Assemble

You still need something to lay clips on a timeline. That part is unavoidable, and it is also the easy part. CapCut, DaVinci Resolve, iMovie, and Premiere all handle it, and the skills involved are cutting and arranging rather than animating.

Assembly order that works: lay down the voiceover or script timing first, drop screen recordings under it, then place graphics at the marks. If a graphic and the narration are fighting for attention, the graphic wins and the narration should go quiet under it for a beat.

### 5\. Cut the variations

The teaser is the strongest 20 seconds of the feature reveal with the narration replaced by on screen text. The hero loop is usually a single graphic or a single interaction, trimmed so the first and last frame match.

For vertical, do not just crop. Re export the graphics at 9:16 so the text stays readable, and let the screen recording sit in a band in the middle with graphics above and below it. A cropped 16:9 chart usually loses its axis labels, which makes it decorative rather than informative.

The same comparison as above, regenerated at 9:16 rather than cropped down to it.

## Specs by placement

| Placement                  | Ratio       | Length            | Audio                   |
| -------------------------- | ----------- | ----------------- | ----------------------- |
| Landing page hero          | 16:9 or 4:3 | 8 to 15s, looping | Muted, must work silent |
| X / LinkedIn               | 16:9 or 1:1 | 15 to 45s         | Captions required       |
| Instagram Reels / TikTok   | 9:16        | 15 to 30s         | Captions required       |
| Product Hunt / launch post | 16:9        | 60 to 120s        | Voiceover or captions   |
| YouTube                    | 16:9        | 60 to 180s        | Voiceover               |

Assume every social placement plays silent on the first pass. If your video only makes sense with sound, it does not make sense.

## What this approach does not solve

Skipping the editor has real costs, and it is worth being clear about them before you commit.

You will not get a cinematic brand film this way. If your launch needs a live action piece with actors, location, and a colourist, hire people. Animated graphics over screen capture is a format, and it suits software products that do something visible on a screen. It suits a physical product launch much less.

You also will not get 3D product renders. If the hero shot is your hardware rotating in space, that is a different pipeline.

And someone still has to write the script. The tooling has removed the animation bottleneck, which means the writing is now the thing that determines whether the video is good. That tends to surprise people.

## The version that ships beats the version that does not

The most common failure mode is scope. A team decides the launch video should be excellent, which makes it a three week project, which means it slips past the launch. A 45 second feature reveal with four clean graphics, published on launch day, does more for you than a beautiful film published two weeks later.

Build the three cuts, publish them, and improve the hero loop later once you can see which part of the product people ask about.

Animated visuals for your videos. In seconds.

Moshion generates the animated visual you need. Describe it, export as MP4, drop it in your editor.

[Try Moshion](/create)