You open Spotlight, type three letters, and wait half a second for it to decide whether you wanted an app, a Wikipedia summary, or a unit conversion you never asked for. That friction is why a chunk of macOS power users replaced Spotlight years ago. The two names that come up are Raycast and Alfred, and in 2026 the gap between them is less about "which can launch an app faster" and more about how you want to extend the thing once it's bound to your hotkey.
We spent time driving both as a daily launcher — the same muscle-memory tasks, the same hotkey, the same goal of not thinking about the tool itself. Here's how they actually diverge.
Two philosophies of extensibility
Alfred has been shipping since 2010, and it shows in the best way: it is bootstrapped, stable, and unapologetically a power-user appliance. Its extension model is the Workflow — a visual node editor where you wire triggers to actions, scripts, and outputs. If you can write a shell, Python, or AppleScript snippet, you can bolt it into a workflow without learning a framework. The result is portable (a .alfredworkflow file is just a bundle) and survives across versions with little drama.
Raycast, which arrived in 2020 and is venture-backed, took the opposite bet. Its extension API is TypeScript and React, distributed through an in-app store with hundreds of published extensions. That means writing one is a real Node project — npm, a build step, a component tree — but it also means extensions render native-feeling list UIs, support real form inputs, and get reviewed before they hit the store. The barrier to author is higher; the barrier to install someone else's is one keystroke.
The practical test: if your customizations are mostly small scripts you already wrote for other purposes, Alfred lets you reuse them with almost no glue. If you want a polished, shareable extension with a UI — a Linear search, a Jira ticket creator, a Homebrew manager — Raycast's store almost certainly already has one, and it'll look like part of the app.
Raycast also ships a lot in the box that Alfred treats as add-ons or leaves to you: clipboard history, snippets, window management, a calendar peek, and an AI command layer are all built in. Alfred keeps its core lean and pushes clipboard history and snippets behind the paid Powerpack, with window management left to a companion tool like Rectangle. Neither approach is wrong — Raycast wants to be a hub, Alfred wants to be a fast, composable primitive.
Pricing, ownership, and the trust question
This is where the two products feel most different, and it's worth being precise rather than hand-wavy.
Alfred is free to launch apps and search. The paid Powerpack — a one-time license, historically in the £30–£60 range depending on whether you buy single or the lifetime "mega" tier — unlocks workflows, clipboard history, snippets, and the rest. You pay once, you own it, and updates within your purchased major version are free. There is no subscription and no account required.
Raycast's core is free and genuinely usable on its own. Raycast Pro is a subscription (roughly $8/month billed annually, more month-to-month) that adds the AI features, cloud sync, unlimited clipboard history, and custom themes. The free tier covers a lot; the moment you want AI built into your launcher or settings that follow you across machines, you're renting.
The ownership distinction matters beyond dollars. Alfred runs entirely local, keeps no account, and sends no telemetry — for people who care about exactly what their always-on launcher is doing, that's a feature. Raycast requires an account for sync and AI, and its roadmap is shaped by the need to eventually justify its funding. Neither has done anything to forfeit trust, but the incentive structures are different, and a launcher is about as privileged a piece of software as you'll run.
A launcher sees every keystroke you type into it and sits in front of your entire system. Before you grant Accessibility and Full Disk Access to either app, check what data leaves the machine. Alfred's local-only default and Raycast's account model are a real decision point, not a footnote.
Whatever you capture — clipboard snippets, scratch notes, links you fire off a workflow to save — needs a durable home, not a buffer that rotates out. A launcher is great at capture and lousy at retrieval a week later.
Which one earns your hotkey
Speed is close enough to call a tie. Both bind to a global hotkey and return results faster than you can finish typing. Alfred has a long reputation for staying light on memory and never getting in its own way; Raycast is native Swift and feels just as instant, though the heavier feature set means it carries more in the background.
The honest decision tree looks like this. Choose Alfred if you want a one-time purchase, a local-only tool with no account, and an extension model that reuses scripts you already have. It rewards people who like to compose small pieces and don't want their launcher to also be a platform. Choose Raycast if you want batteries included — window management, clipboard, snippets, and AI in one surface — and you value a store full of polished, install-in-one-click extensions over writing your own glue.
For developers specifically, Raycast's edge is the extension ecosystem: there's a good chance the tool you use (GitHub, Linear, Vercel, your password manager) already has a maintained extension that turns a multi-click task into a two-word command. Alfred's edge is that nothing is hidden behind a service, and a workflow you build today will still be a portable file in five years.
There's no universal winner here, and anyone who declares one is selling you their own workflow. Alfred is the better fit for the script-composing, own-it-once, local-only crowd. Raycast is the better fit for people who want a maintained ecosystem and don't mind a subscription for the AI layer. Pick the incentive structure you trust and the extension model that matches how you actually work — then stop thinking about the launcher, which is the entire point of having one.
Originally published at pickuma.com. Subscribe to the RSS or follow @pickuma.bsky.social for new reviews.
Top comments (0)