How Raycast announces features to developers — and how to copy it
A teardown of Raycast's lightweight feature-clip format: native chrome, real keystrokes, no narration, and why the same format works for community extensions.
Raycast ships features constantly, and a large share of what people use it for comes from a community of extension developers rather than the core team alone. Its feature videos are built to match that volume: a format so lightweight it can be applied to a minor extension update with the same care as a headline feature, without either looking under- or over-produced.
Everything below is from watching the published feature and extension videos. Nothing about Raycast's internal production process is public, so the cost section is an estimate with its assumptions stated.
1. What they publish, and how often
Short clips for individual features and notable community extensions, at a cadence closer to a running feed than a launch calendar — frequent, small, and treated as routine documentation rather than an event each time.
2. The mechanics, named
| Element | What they do | Why it works |
|---|---|---|
| Chrome | Real native macOS window chrome, unmodified | The audience is verifying this runs on their actual machine, not watching a stylised recreation |
| Interaction | Real keyboard shortcuts, shown at natural speed | The shortcut itself is often the feature; showing it any other way would misrepresent it |
| Scale | 1:1, uncropped — the launcher shown at its real on-screen size | Raycast's compact size is part of its pitch; cropping in would contradict that |
| Narration | None | The audience already knows the product; the clip only needs to show what changed |
| Treatment of extensions | Same format for community and first-party features alike | Signals the ecosystem is taken seriously rather than treated as a lesser product tier |
The clip format is cheap on purpose — cheap enough to be worth making for a feature that will only ever get thirty seconds of anyone's attention.
3. What something like this costs
Stated as an estimate: given the near-total absence of post-production — no voiceover, no music, minimal editing beyond a trim — a clip in this style typically costs $200–$600 from a freelance editor, or under an hour in-house once a recording template exists. This is close to the floor of what a produced feature clip can cost, which is precisely what makes the volume possible.
The only real cost that scales with volume is triage: deciding which of many small features and extensions are worth a clip at all. That is an editorial decision, not a production one, and it is the actual bottleneck behind a format this cheap to execute.
4. The version you can make this week
A feature clip cheap enough to make for everything
- Record the feature exactly as triggeredThe real keystroke or click sequence a user would actually perform — not a cleaned-up or slowed version of it.
- Keep native chrome visibleDon't crop the window or hide the OS frame. For a utility or developer tool, the native chrome is part of the credibility.
- Cut the instant the result appearsNo lingering, no replay. This format is fast because it trusts the viewer to understand a result the moment they see it.
- Skip narration by defaultAdd a caption only if the feature's name or trigger genuinely isn't obvious from the footage alone.
- Use the same format for every feature, big or smallConsistency of format, not scale of production, is what makes a high volume of clips sustainable and what signals that even small updates are taken seriously.
What not to copy
Do not drop narration or context for an audience that does not already know your product — this format's speed depends entirely on the viewer already understanding what they are looking at. A first-time visitor watching an unexplained keyboard shortcut fly past will learn nothing from it. Use this style once an audience exists that can read it; use a slower, explained format to build that audience in the first place.
Questions people ask
How does Raycast make videos for so many small features?
By using one lightweight, repeatable format — a short screen recording with native macOS chrome, real keystrokes, and no voiceover — so a clip costs little enough in production time to justify making one for even a minor feature or extension.
Why don't Raycast's feature clips have voiceover or music?
The audience is developers and power users already familiar with the product category, so the clip's job is to show exactly how a feature is triggered and what it does — narration would add time without adding clarity for an audience that reads a keyboard shortcut faster than a sentence explaining it.
Does treating community extensions the same as first-party features risk looking unpolished?
The opposite, for this audience: giving a community extension the same clip format as an official feature signals that the platform takes its ecosystem seriously, which matters more to a developer audience than uniform production values would.
Can a consumer product use this no-narration, native-chrome style?
Only where the audience already understands the product and the interaction is genuinely simple to follow without commentary. A first-time user encountering an unfamiliar interface usually needs the context a voiceover or caption provides — this style trades accessibility for speed, deliberately.
Keep reading
How Vercel ships a video with almost every release — and how to copy it
Fewer words, more terminal, almost every release. Vercel's ship videos are built for a cadence most companies could actually sustain — if they made the same trade-offs.
How Figma makes a launch feel like an event — and how to copy it
A pause, then the product. Figma's launches borrow keynote pacing rather than tutorial pacing — and the pause is the only part that actually matters.
The six videos every software company needs
Most teams start by picking a video tool. Start by naming the job instead: these six cover almost every request a software company gets.