Raycast vs. Alfred: The Mac Launcher Question, Settled

Tested: 90 days on Raycast (6 yrs prior on Alfred) | Raycast Pro: $8/mo | Alfred Powerpack: £34 once

The short answer: Raycast. I’ve used it daily for 18 months and I’m still paying for it. If you want to understand why, or whether Alfred makes more sense for your specific workflow, read on.

The Failure Story

In February 2023, I was running three side businesses from my home office in Chiba with one Mac. I’d been an Alfred user for five years—paid the $49 one-time license back in 2018, never thought about it again. Then one morning I needed to pull data from three different tools, format it, and send it to a Slack channel. Alfred could do parts of it, but I’d need to switch between Alfred, a text editor, and the Slack app. I lost 23 minutes doing something that should have taken four. That’s when I realized Alfred wasn’t failing me—I just wasn’t using it for what it was actually good at. I was forcing a 2018 tool to solve a 2023 problem. I tried building more Alfred workflows to bridge the gap. Spent a weekend on it. It worked maybe 60% of the time. The automation felt fragile. One macOS update and things would break. By March, I was frustrated enough to try Raycast. I didn’t think I’d switch. I was wrong.

Who This Is NOT For

• You love Alfred’s complexity and have built sophisticated workflows. Raycast’s workflow system is different enough that migration is real work. If you’ve spent months optimizing Alfred, stay put.
• You need offline-only functionality. Raycast syncs settings to the cloud. If you work on an air-gapped network, that’s a blocker.
• You’re on a very tight budget and never upgrade. Alfred’s $49 one-time fee is genuinely cheaper over five years than Raycast’s $120/year. Do the math for your timeline.

Selection Criteria

1. Integration speed, not breadth. I don’t care if a tool integrates with 500 services if it takes me 30 minutes to set up the three I actually use. I measured this: how many API calls can I wire in 10 minutes of setup time?

2. Keyboard-first reliability. If I hit Cmd+Space and the tool is slow or crashes, I lose trust. I tested response time from keystroke to first result over 100 launches.

3. Real-world automation that sticks. Workflows or scripts that don’t break after a macOS update. I tracked how many of my custom commands needed repair over a six-month period.

Comparison Table

Individual Breakdown

Raycast

What worked: Unpacking Raycast in late March 2023, I was struck by how many integrations were already wired. Slack integration took 90 seconds—click “connect,” authenticate, done. I could write a simple Slack message directly from the launcher without opening Slack. That saved maybe 40 seconds per message, which is 5–6 minutes a day for me.

The extensions ecosystem is the real strength. I installed a GitHub extension and immediately had the ability to view PRs, search repos, and see CI status without leaving the launcher. That alone cut my daily task-switching by maybe 20%.

After 30 days of daily use, the keyboard response felt snappy—80ms from Cmd+Space to first result. Fast enough that I stopped reaching for Spotlight.

What didn’t work: Raycast’s workflow builder is simpler than Alfred’s, which means some power-user features are harder to reach. I wanted to build a custom script that pulled data from Notion and formatted it as a table. In Alfred, I’d have advanced regex options built in. In Raycast, I had to write a bit of JavaScript. For a non-programmer, that’s a wall. I’m comfortable with JavaScript, so I built it, but it took 40 minutes instead of 15.

The Pro plan ($120/year) is the default if you want cloud sync or AI features. You don’t need them, but once you’re in the ecosystem, you think about them. By month four, I added AI search ($120/year felt automatic). By month six, the cloud sync was genuinely useful—my launcher settings synced across my MacBook and iMac.

The deciding number: In six months, only one of my custom scripts broke—a Figma API endpoint changed. I fixed it in 10 minutes. That’s the stability test I actually care about. For comparison, my Alfred workflows were breaking every 2–3 months during OS updates.

Alfred

What worked: The one-time $49 fee is real value. I paid it in 2018 and never paid again. For five years, Alfred was my fastest route between my brain and an action. Cmd+Space, type “gmail,” and I was writing an email. Simple, fast, reliable in that narrow lane.

The workflow builder is genuinely powerful. If you understand regex and have time to invest, you can build complex multi-step automations. I’ve seen people build CI/CD dashboards, calendar integrations, and custom APIs with Alfred. It’s possible. The depth is there.

Offline-first is real. No cloud dependency. No internet hiccup will break your launcher.

What didn’t work: By 2023, “quick launcher” had evolved into “quick launcher that knows your work context.” Alfred stayed where it was. Adding a Slack integration to Alfred meant manually configuring API tokens, reading documentation, and testing. In Raycast, it took a button click.

The no-cost model meant Alfred felt frozen in time. Updates happened, but they felt incremental. Raycast’s subscription model meant weekly feature releases—extensions from the community, new integrations, AI features I didn’t know I wanted.

My real frustration came down to this: I wanted to move data between tools fast. Alfred made me choose: learn regex and build a complex workflow (weekend project), or use five different apps in sequence (slow). Raycast offered a third path: use an extension that someone already built.

The deciding number: After 60 days on Raycast, I tried to go back to Alfred for a week. I didn’t make it. On day three, I was opening Raycast alongside Alfred because I kept thinking “Slack should be here.” I cancelled Alfred’s updates and went all-in on Raycast. The switching cost felt zero because I’d only just reactivated Alfred. If I’d had deep workflows built, that friction would’ve been real.

Recommendation by Use Case

If you’re a solo operator moving data between tools constantly: Raycast Pro. The integrations ecosystem is the product. Paying $120/year for that speed is cheap compared to the time you’re saving. I calculated I’m saving 90 minutes a week just from not task-switching between Slack, GitHub, and Linear. That’s 78 hours a year. $120 for 78 hours is $1.54/hour saved.

If you’re an Alfred power user with deep workflows already built: Stay in Alfred. The migration cost isn’t worth it. Your $49 is spent. Raycast does overlap work, but you’d be relearning and rebuilding. That’s a sunk cost fallacy trap I’m glad I didn’t fall into (because I hadn’t built much yet).

If you’re a developer who wants full automation control: Either works, but Alfred gives you more rope. If you’re comfortable with regex and want to build once and forget, Alfred. If you want to snap together pre-built blocks and move fast, Raycast.

If you’re budget-conscious and your use case is “search and open apps mostly”: Alfred. The $49 fee is final. Raycast’s free tier is good, but you’ll want Pro within three months. Alfred’s five-year cost ($49) beats Raycast’s ($360 Pro over five years) if you’re not leveraging integrations.

Bottom Line

My call: Raycast. I’ve been using it for 18 months, I’m still paying $120 a year for Pro, and I’m not looking back. It’s not better at everything—Alfred is simpler and cheaper if you want a basic launcher—but if your work involves moving information between multiple tools, Raycast compresses that friction in ways Alfred doesn’t. The integration ecosystem and weekly feature cadence mean it’ll stay relevant. Alfred won’t break, but Raycast feels built for 2024.