Bolt builds a working booking app in about 7 minutes
Verdict
Bolt.new generated a working version of SlotKeeper, a booking and waitlist app for a demo yoga studio (Maple Yoga Studio), in about 7 minutes with zero fix prompts and zero clarifying questions. It passed all 10 checklist items, published free in-tool, and finished on the free plan for $0 (the account had a 1M-token bonus for completing a survey). We found 1 major bug (slot times shift with the viewer's timezone) and 2 minor ones. Score: 94.75/100. Bolt and Base44 are effectively tied (94.75 vs 93.5); the gap is within one reviewer's design judgment.
| Score box | |
|---|---|
| Overall | 94.75/100 |
| Works end to end | 30/30 (10/10 pass) |
| Time to first working version | 15/15 (~7 min) |
| Bugs | 11/15 (1 major, 2 minor) |
| Deploy ease | 15/15 (in-tool Publish, free bolt.host URL) |
| Cost | 10/10 (free plan, $0; 1M-token survey bonus on the account) |
| Design / UX | 13.75/15 (polish 5, mobile 5, clarity 4; raw 14; one reviewer) |
| Tier used | Free (plus a 1M-token survey bonus) |
Try Bolt →
The Bolt-built booking page on desktop. Note the 2:00 AM and 3:00 AM slot times, which come from the timezone bug below.
What we tested
SlotKeeper: customers book slots and join a waitlist, and an owner manages slots and bookings. We tested it on Oct 21, 2026, on Bolt.new's free plan. The account received a 1M-token bonus for completing a survey. Our method: one fixed prompt, at most 5 fix prompts, no hand-editing of code, a 90-minute cap, a 10-item checklist and a fixed rubric. Full method
Live app: https://slotkeeper-booking-w-wqsh.bolt.host
How it went
We submitted the prompt at 10:29:30 EDT, and generation finished at about 10:36, so roughly 7 minutes. Bolt asked no clarifying questions and needed no fix prompts. It reported "The site is ready to deploy" and seeded 3 slots.
The stack is Bolt Database (Postgres with row-level security) and a frontend with hash routing.
Publishing was the in-tool Publish button and then Publish in the dialog. That gave us a free, public *.bolt.host URL with no paywall.
Our test box ran out of memory twice (on the first send and when loading the preview), which delayed our verification. Those crashes were on our side, not Bolt's, and aren't counted against it.
Checklist: 1. 3 seeded slots: pass 2. Booking reduces spots left, with confirmation: pass 3. Form validation: pass 4. Bookings persist after refresh: pass 5. Waitlist with position #1: pass 6. Duplicate email blocked: pass 7. Owner sign-up and login, and logged-out /admin redirects: pass (via the app's /#/admin route; plain /admin is logged as a bug) 8. Create, edit and delete a slot: pass 9. Cancelling auto-promotes waitlist #1: pass 10. Stats line correct (3/2/1 → 3/2/0): pass
The booking confirmation screen.
The owner's admin dashboard, with the stats line and slot management.
What broke
1 major and 2 minor bugs: 1. Plain /admin doesn't redirect (minor). The app uses hash routing, so the admin page lives at /#/admin. Logged out, /#/admin correctly redirects to login, but visiting plain /admin just shows the landing page. To reproduce, open the live URL with /admin on the end while logged out. 2. Slot times shift with the viewer's timezone (major). A booker can see the wrong time for a session, which we count as major under our rubric (like a wrong spot count). The same slot showed 2:00 AM in a regular browser and 5:00 AM in a headless one, and the seeded times look odd. To reproduce, open the landing page in browsers with different timezone settings and compare the slot times. 3. A brief "0 classes" loading flash (minor). For about 1 second on load, the page shows "0 classes" before the sessions appear. To reproduce, hard-refresh the landing page.
Mobile view (headless browser). The times differ from the desktop shot (5:00 AM vs 2:00 AM for the same slot) because of bug 2.
Caveats
- Design is scored by one reviewer. 13.75/15 (polish 5, mobile 5, clarity 4) is one reviewer's judgment, the same method we used for Base44. The 1.25-point gap between Bolt and Base44 is within that judgment.
- Owner sign-up is open to anyone, with no email verification. The prompt asked for owner sign-up, so we didn't score this. But we didn't run a second-owner data-isolation test, so we can't say whether one owner can see another's data.
- We didn't check whether database keys are exposed in the client.
- The free-tier result comes with a bonus. The account received a 1M-token bonus for completing a survey, and Bolt didn't show a token counter, so we can't say whether a standard free allowance would have been enough.
- Data persistence across a redeploy wasn't tested.
How Bolt compares
| Tool | Score | Time to first working version | Deploy | Critical bugs | Minor bugs | Tier |
|---|---|---|---|---|---|---|
| Bolt | 94.75/100 | ~7 min | In-tool, free | 0 (1 major) | 2 | Free (+1M-token survey bonus) |
| Base44 | 93.5/100 | 8–9 min | In-tool, 4 clicks, free | 0 | 4 | Free |
| Lovable | Test in progress | |||||
| Cursor | Test in progress |
Pros and cons
Pros - Working app in about 7 minutes with zero fix prompts - Free in-tool publish to a public URL - Finished on the free plan for $0 (with a 1M-token survey bonus) - Real Postgres database with row-level security - No critical bugs
Cons - Major bug: slot times shift with the viewer's timezone - Hash routing means plain /admin doesn't redirect - Open owner sign-up with no email verification - No visible token usage
Who it's for
Non-developers who want a working booking site with an owner admin, built and published free (our run used a 1M-token survey bonus). Before you take real bookings, fix the timezone handling and lock down owner sign-up.
Build your booking app with Bolt →
FAQ
Can Bolt build a booking app? Yes. In BenchCoded's SlotKeeper test, Bolt.new built a booking and waitlist app that passed all 10 checklist items. One related bug: because the app uses hash routing, plain /admin shows the landing page instead of redirecting to login (the admin page itself, /#/admin, redirects correctly).
How long does it take Bolt to build a booking app? In our test, Bolt.new finished generating a working booking app in about 7 minutes, with zero fix prompts and zero clarifying questions.
What bugs did Bolt's booking app have? We found 1 major and 2 minor bugs. Major: slot times shift with the viewer's timezone, so a booker can see the wrong time for a session. Minor: plain /admin shows the landing page instead of redirecting to login, and the page briefly shows "0 classes" for about a second while loading.
What score did Bolt get? Bolt.new scored 94.75/100 on BenchCoded's SlotKeeper benchmark: 30/30 works end to end, 15/15 time, 11/15 bugs, 13.75/15 design (scored by one reviewer), 15/15 deploy and 10/10 cost.
Is Bolt free for building a booking app? Yes, in our test. Bolt.new built and published the app on its free plan for $0 (the account had a 1M-token bonus for completing a survey), on a free bolt.host URL. Bolt didn't show a token counter, so we can't say whether a standard free allowance without that bonus would have been enough.
Do I need to code to use Bolt? Not for our SlotKeeper build. Our method bans hand-editing code, and Bolt.new produced a working app from one prompt with zero fix prompts.