Launch films · landscape 16:9
How to make a product launch video
A product launch video works best as a ~60-second problem-first film: name the pain in the first five seconds, show the product solving it in real UI, then close on the promise and a call to action. The hard part is not the script — it is producing UI motion that looks designed rather than recorded. With baz, you point the Launch video skill at your product (repo, live site, screenshots) and the agent recreates your interface as animated components, writes the script from your own marketing copy, and choreographs voiceover, music and sound effects into a finished cut.
The shortcut: run a skill instead of starting from scratch
A skill is a context file that teaches an AI agent one specific video style — its structure, pacing, typography, motion grammar and sound design. Hand an agent the skill plus your product, and it rebuilds that video for you. The steps below are what the Launch video skill already encodes, so you can follow them by hand or let the skill carry them.
baz is a library of video skills: context files that teach an AI agent a specific motion-graphics style, so it can rebuild that video for your product, with your brand, from your inputs. Run a skill inside your own coding agent — Claude Code, Codex — or directly in the baz editor.
What you need before you start
Required
- A live product URL, or the repo — the skill reads brand tokens, logo and real UI labels from it
- Screenshots of your main product surface, plus one or two secondary screens
- Your brand assets: logo file and the colours you actually ship
Makes it materially better
- Customer logos for a social-proof wall
- One hard metric you are allowed to claim on screen
- Your tagline and the exact CTA URL you want to end on
- A feature list, so the middle act has real substance instead of adjectives
How to make a product launch video: step by step
Decide the one problem the video is about
Launch videos fail by listing features. Pick the single painful thing your product removes and commit the first five seconds to it. Everything after that is evidence. If you cannot say the problem in one sentence, the video will not hold for sixty seconds.
Gather the evidence before you write anything
Collect the product URL, screenshots of the surfaces you want on screen, your logo, and your brand colours. The skill can fall back to your landing page, GitHub README or App Store listing, but the output is measurably better when you hand it the real assets rather than making it infer them.
Let the skill write the script from your own copy
Point the agent at your product and let it draft a problem-first script using your live marketing language. This matters: a script written from your own site sounds like you, and it keeps the claims to things you already say publicly. Read the draft and cut anything you cannot defend.
Recreate the UI instead of recording it
Screen recordings age badly, capture at the wrong aspect ratio, and show whatever state your account happened to be in. The skill rebuilds your interface as animated components — real labels, real layout, real brand colours — so the camera can push into a single button or scale a whole dashboard without going soft.
Choreograph the camera, then the sound
Add the macro moves — push-ins on the moment that matters, a pull-back to show scale. Then layer voiceover, a music bed, and sound effects tied to what is actually mutating on screen. Sound sold to the picture is the difference between "made with AI" and "made by a studio".
Review, revise the parts that are wrong, and export
Watch it once at full speed without stopping. Note only what actually bothers you, then ask for those specific changes — re-time a beat, swap a screenshot, rewrite one line. Rebuild only what changed, then export at 1080p for the launch post.
What the finished video actually contains
Approximate timings for a ~60 seconds cut. Treat them as a shape to hit, not a spec — the exact durations shift with your script and your music.
| Time | Beat | What is on screen |
|---|---|---|
| 0:00–0:05 | The problem | One sentence, stated plainly over type. No logo yet, no product yet — you have not earned the reveal. |
| 0:05–0:12 | The turn | The product enters as the answer to the sentence you just said. First real UI on screen. |
| 0:12–0:38 | Proof chapters | Two to four capabilities, each shown working in recreated UI with a macro camera move, not described in a bullet. |
| 0:38–0:50 | Scale and credibility | Customer logos, a metric you can defend, or the integrations wall — the beat that says other people already trust this. |
| 0:50–1:00 | Lockup and CTA | Logo resolve, the tagline, and one unambiguous next action. One CTA, not three. |
Mistakes that cost you the view
Opening on your logo
Do this instead: Open on the problem. Nobody has a reason to care about your mark in second one; earn the logo by the end.
Using a raw screen recording
Do this instead: Recreate the UI as components. Recordings lock you to one resolution and one account state, and you cannot push the camera into them without mush.
Cramming in every feature
Do this instead: Two to four capabilities, each properly demonstrated. A launch video is a trailer, not a manual — the point is to make people click, not to make them fully informed.
Writing the script in marketing voice
Do this instead: Say what it does in the words a user would use. Adjectives ("seamless", "powerful") occupy time without transferring information.
Music that ignores the cuts
Do this instead: Put the scene boundaries on the beat. Cuts that land off the music read as amateur even when every individual shot is good.
Run it from your own agent
Skills are designed to run inside a terminal agent like Claude Code or Codex, against a project you control. The same skill produces the same style whether it runs in your agent or in the baz editor.
npm install -g bazaar.it@latest
# Install this skill into your agent (Claude Code, Codex or Cursor)
baz skill add launch-video
baz project create --name "Launch video for my product"
baz prompt "make a product launch video" --stream-json
baz review --summary --json
baz export start --wait --format mp4 --jsonOperations draw from a prepaid balance. See the CLI reference for the full command set.
Common questions
How long should a product launch video be?
Around 60 seconds for the main launch asset. That is long enough for a problem, two or three proofs and a close, and short enough to survive an autoplay feed. Cut a 15-second version from the same project for paid social.
Do I need a voiceover?
Not necessarily. Voiceover carries a launch film well when you have a real argument to make. If your product is visually self-evident, a music-cut version with kinetic type often lands harder — the Supahub Ad skill is built entirely without narration.
Can it read my private repo?
The skill reads whatever you give the agent access to. Run it in your own coding agent against a local checkout and nothing leaves your machine except the assets you choose to upload for the render.
Do I need video editing experience?
No. The skill carries the editing decisions — shot order, pacing, type treatment, transitions, sound design. You supply the raw material and the facts about your product; the agent assembles the cut. You review it and ask for changes in plain language.
Can I run this in my own coding agent instead of the baz editor?
Yes. Skills are designed to be run from a terminal agent like Claude Code, Codex or Cursor. Install the CLI with `npm install -g bazaar.it@latest`, then `baz skill add <slug>` to pull the skill into your agent. The same skill produces the same style whether it runs in your agent or in the baz editor.
What does it cost?
baz bills against a prepaid dollar balance rather than a seat subscription — you top the balance up and each operation (generation, voiceover, render, export) draws a listed amount from it. There is no per-seat fee to make a video, and the price of each operation is published.
Can I change the video after it renders?
Yes, and this is the point of doing it as code rather than as a flat render. Every scene is a real component you can edit — ask the agent to re-time a beat, swap a screenshot, rewrite a line, or change the accent colour, and only that part rebuilds. You are not re-rendering from scratch each time.
Will every video from this skill look the same?
The grammar is the same — that is what makes it a skill and not a template. The content is not: your brand colours, typography, screenshots, copy and data all come from your inputs. Reusing one skill across a series is a feature, because episode three lands in the same visual language as episode one instead of being rebuilt from scratch.
Related guides
Pick a skill. Get the video.
Run the Launch video skill in your own coding agent, or in the baz editor. Usage comes out of a balance you top up, with a listed price per operation — no seat subscription to make a video.