How Vercel ships a video with almost every release — and how to copy it
A teardown of Vercel's high-frequency ship videos: the low-narration, terminal-led style, what sustaining that cadence costs, and the version you can make this week.
Where a lot of B2B software treats a launch video as an occasional, considered production, Vercel's ship videos are built to be made often — close to every meaningful release rather than a handful a year. What makes that sustainable is not a bigger team; it is a format with almost nothing in it to slow the film down.
Everything below is from watching the published ship and feature videos. Nothing about Vercel's internal production process or team size is public, so the cost section is an estimate with its assumptions stated.
1. What they publish, and how often
Short clips tied closely to shipping cadence — a new CLI command, a framework feature, a dashboard change — released frequently rather than batched into quarterly launch events. The frequency is the point: a developer audience learns to expect a film with almost every notable change.
2. The mechanics, named
| Element | What they do | Why it works |
|---|---|---|
| Subject | Real terminal and code-editor footage | The audience trusts a genuine command line over a stylised recreation of one |
| Narration | Minimal or absent — often just captions or none at all | Developers read a screen faster than they process spoken explanation; narration would slow the film down |
| Pacing | Held shots on output long enough to actually read | A cut that arrives before the terminal text can be read defeats the point of showing it |
| Length | Often under 20 seconds | Matched to one change, published at a cadence a longer format could not sustain |
| Typography | Monospace, minimal styling, close to the real editor theme | Nothing about the presentation contradicts 'this is really happening in your terminal' |
The audience reads a diff faster than it processes a voiceover. Silence is a pacing decision, not an absence.
3. What something like this costs
Stated as an estimate: because the format deliberately strips out narration, music design and heavy editing, a single ship video like this is inexpensive relative to most product film — commissioned from a freelancer, likely $300–$1,000 per clip given how little post-production the style calls for. In-house, it is closer to one to three hours: record the terminal, trim, add a caption, publish.
The real cost is not any single video — it is the discipline of shipping one almost every time, which is a process commitment more than a production one. A team that can capture and publish a 15-second clip within a day of shipping is doing the expensive part; the clip itself is close to free.
4. The version you can make this week
A ship video in under an hour
- Record the real command or changeScreen-record the actual terminal session or code diff, not a recreation. Authenticity is most of the credibility this style relies on.
- Trim to the moment that mattersCut everything before the command runs and everything after the result appears. The film is the moment of change, not the setup.
- Hold the output a beat longer than feels naturalThe instinct is to cut as soon as the result appears. Hold an extra second — the audience needs time to actually read it, not just see that something happened.
- Skip the voiceoverTry it with no narration first. Add a single caption only if the change genuinely needs a name attached to it.
- Publish it smallThis format does not need a launch page or a campaign. Post it where the audience already is — a changelog, a developer community, a single tweet.
What not to copy
Do not drop narration entirely if your audience is not developers — the silence works here because the viewer can read code and terminal output as fast as speech; a non-technical audience watching a silent UI change usually needs a caption or a voice to know what they are looking at. Copy the discipline of a small, fast, repeatable format matched to cadence — not the specific absence of a voice track.
Questions people ask
How does Vercel make a new video for almost every feature?
By keeping each film small and fast to produce rather than treating every ship as a major production: real terminal or editor footage, minimal or no voiceover, and a consistent short structure that a small team can repeat weekly rather than quarterly.
Why do developer-audience ship videos use less narration than consumer product videos?
Because the audience reads code and terminal output faster than it processes spoken explanation, and a voiceover narrating what is already legible on screen slows the video down for exactly the people it is for. Silence, or a single caption, respects that the audience is technical.
Can a non-developer product copy this low-narration style?
Only if the audience can genuinely read the screen as fast as a developer reads code — for most consumer or business software, some narration or captioning is still doing real work. The transferable lesson is cadence and restraint, not the absence of a voice track.
What's the trade-off in shipping a video for almost every feature?
Frequency costs distinctiveness. If every ship gets a film, no single film feels like a major event — which is the correct trade for a fast-moving developer product, and the wrong one for a company that wants a handful of releases each year to feel unmissable.
Keep reading
How Linear makes its launch videos — and how to copy it
No presenters, no stock, no explaining. The most copyable film style in B2B software, taken apart — and rebuilt at a fraction of the effort.
How Stripe films something you cannot see — and how to copy it
Payment infrastructure has no interface. Stripe's videos solve that by drawing one — and the technique generalises to any product whose real work happens off screen.
Product demo video examples, annotated
Four finished films, and the one decision behind each that's worth copying — rather than a highlight reel with no explanation of why any of it works.