How to Make a Product Launch Video on a Mac (Founder-Led, One Recording)

A product launch video has one job: a stranger who has never heard of your product watches it once and understands what it does and why they should care. Most of the results for this search are agency showreels and template sites. This guide is for the founder who is going to make it alone, on a Mac, in an afternoon.

The short version:

• Length: 60 to 90 seconds for the main cut.
• Content: one problem, the product solving it from start to finish, the result on screen.
• Method: record your screen and your camera together, once, while you click through the product. Fix the start and the end, not the middle.
• Versions: a widescreen cut for YouTube and your launch post, a vertical cut for X and LinkedIn, and a silent loop for your homepage, all from the same recording.

What a launch video is (and what it isn’t)

Three kinds of product video get mixed up, and each needs a different script.

The launch video is the only one of the three that has to work on someone with no context and no patience. So it shows one scenario, the one that makes your ideal user think “I need that”, and nothing else. Every other feature waits for the demo and the announcements that follow.

If your product is screen software, you don’t need animation for this. A real recording of the real product is more convincing than a drawing of it, and it’s the same reason an explainer can work without animation.

The 60 to 90 second structure

Write to this timing before you open anything. A launch video that runs long almost always has no script.

Three things to cut:

• Logo intros. The first five seconds decide whether anyone stays. Spend them on the problem.
• Feature lists. Ten features in 45 seconds reads as none. Show one job.
• “Hi, I’m the founder of…” Say who you are in the close, if at all. People care about what the product does for them first.

A script for this is about 150 to 220 words. Read it aloud with a timer before you record. If it runs past 90 seconds, cut a sentence from the middle, not the hook.

Before you record

Ten minutes of setup decides whether you need one take or six.

• Use a demo account with believable data. Real-looking names and numbers, no “Test 1”, no notification badges.
• Pick one window and one size. Record the product window, not the whole display. A window around 1440 x 900 points is readable after export.
• Clean the screen. Turn on Do Not Disturb and hide what you don’t want seen with this clean-screen checklist for Mac.
• Put the script where you can read it. In CursorClip, open Notes from the recording controls and paste the script. It’s the built-in teleprompter: it sits beside the product while you talk and stays out of the recording.
• Check the microphone. A plug-in USB mic close to your mouth sounds far better than the laptop’s built-in one. Record ten seconds and play it back.
• Rehearse the clicks twice. Any step where you hesitate in rehearsal is a step to simplify.

Record it once

This is where a launch video usually turns into a two-day editing job: you record the screen, then add zooms by hand so viewers can read the small buttons, then try to fit your face in. CursorClip does the zooms and the camera while you record, so the editing that’s left is the start and the end.

1. Open CursorClip, pick the product window, and turn on your camera and microphone.
2. Start recording and look at the camera for the hook. Say your first line before you touch the product.
3. Click through the one job at about two thirds of your normal speed. CursorClip places zooms on your clicks, so the viewer always sees the part you’re working on, at a size they can read. Zooms follow clicks only, not hovering, so click where you want people to look.
4. Pause for a second before each important click and after each result. The pause gives the zoom time to settle and the viewer time to read.
5. Hold the final result on screen for two or three seconds, look back at the camera, and say the close. Then stop.

If you fluff a line, pause, and say it again. You’ll cut the bad take in the next step.

Edit: the start, the end and the frame

Open the recording in the editor. Three passes are enough.

1. Cut. Use Split on the timeline to remove the false start and any fluffed lines. If you narrated, Find pauses in the Cut panel shows the long silences on the timeline first, and nothing is removed until you press Remove.

2. Frame with layouts. CursorClip 2.1.0 has 20 layouts for a screen and camera recording. For a launch video, a simple pattern works:

• Hook and close: Full camera (in the Camera only group), so it’s you talking to the viewer.
• Product part: a Camera bubble layout such as Corner bubble or Circle, or Side by side if your face matters to the story.

Move the playhead to where the frame should change and use Change layout from here in the Layouts panel. You don’t record the parts separately: it’s one recording, framed differently in each part.

3. Check the zooms. Play it back once. If a zoom lands somewhere you don’t want, move, trim or delete it, or add one where you need it. How automatic zooms work covers the controls.

Watch the whole thing once more at full size and once on your phone. If you can’t read the product on the phone, the zoom is too wide or the window was too big.

Export one recording, several versions

You don’t need a new recording for each place the video goes. Change the shape in the editor and export again.

Which file format. Export’s Format setting defaults to Automatic. Up to about 2.3 million pixels (1920 x 1200), which includes 1080p widescreen and the 1080 x 1920 vertical cut, Automatic makes H.264, which plays everywhere. Above that size (4K, for example) it makes HEVC on Apple silicon, which exports faster and is about a third smaller. YouTube lists HEVC (H.265) among its supported upload formats. If any site refuses a file, export again with Format set to H.264.

Captions. Most feed video plays muted, so captions help. CursorClip 2.1.0 doesn’t make them; captions arrive in the next release. Until then, keep the key words on screen in your product (a headline in the UI, a clearly named button), or add captions in another tool after you export.

The version for your homepage

The top of your homepage needs a different cut of the same recording: a silent loop that plays on its own, plus the full narrated video for anyone who wants it. Most browsers autoplay a video only when it is muted, so a narrated video at the top of a page usually sits there as a still frame until someone clicks.

Our own homepage does this, with a shorter loop. Checked on 27 October 2026, the hero loop is 15.2 seconds, 1600 pixels wide, 30 fps, H.264 with no audio track, and about 1.1 MB. Click it and the full demo plays: 39 seconds with sound, about 2.9 MB. The loop is a CursorClip auto-zoom recording.

Make the loop (20 to 30 seconds):

1. Export the main versions first, then keep working in the same project.
2. Trim to the product part only: the clicks that show the one job, ending on the result. Give it a Screen only layout, such as Screen on background, so no camera shows.
3. In the Audio panel, turn off Include Microphone and Include System Audio.
4. Export MP4 at 1080p or smaller. Automatic gives you H.264 at that size.

Put it on the page:

<video autoplay muted loop playsinline preload="metadata" poster="/launch-loop-poster.jpg">
  <source src="/launch-loop.mp4" type="video/mp4" />
</video>

muted and playsinline are what let it autoplay on phones. The poster image is shown until the video loads, so use the loop’s first frame. Link or open the full narrated version from a button or a click on the loop.

Keep the file small. File size is bitrate times length: a 25-second loop at 1 Mbps is about 3 MB, and at 0.6 Mbps (what our hero loop runs at) about 2 MB. A product UI with little motion compresses well. If the export is much bigger than that, use the video bitrate calculator to pick a target and the video compressor to get there, then check that small text is still sharp.

When to hire an agency instead

Doing it yourself is the right call when the product has a screen, you’re the person who explains it best, and you’ll need to remake the video when the product changes. An agency earns its fee when:

• the product has no screen to show (hardware, a service, an API with no UI);
• the story needs animation, actors, locations or a brand film;
• the launch is big enough that the video is an ad campaign, not a product walkthrough.

Even then, a founder-made recording is a good brief for the agency, because it shows them exactly what the product does.

Launch video checklist

Before you record:

• Script of 150 to 220 words, read aloud under 90 seconds
• One job chosen, hook written as one sentence
• Demo account with believable data
• Product window sized, Do Not Disturb on, screen cleaned
• Script pasted into Notes, mic checked with a ten-second test

After you record:

• False start and fluffed lines cut
• Full camera for the hook and close, screen with camera for the product part
• Zooms checked once at full size and once on a phone
• Widescreen export for YouTube and your launch post
• Vertical 1080 x 1920 export, trimmed to 30 to 45 seconds, for X and LinkedIn
• Silent 20 to 30 second loop for the homepage, with the narrated version a click away

If your launch also comes with new features for existing users, follow it with a short feature announcement video for each one. For the tool side of all this, see the product demo video maker for Mac.