We sent the same prompt to all four: build a multi-user to-do app with login, a database and a public URL. Then an agent ran 10 feature checks, tried to read one user's data as another user, and checked it on a phone-sized screen.
| Tool | Score | Prompt → live | Fix prompts | Free-plan cost |
|---|---|---|---|---|
| Lovable | 100/100 · tier pending v2 | under 4 min | 0 | 2.10 credits (free plan) |
| Bolt | 100/100 · tier pending v2 | about 6 min | 0 | about 101K of 1M monthly tokens (free plan) |
| Base44 | 100/100 · tier pending v2 | about 3.5 min | 0 | 1 of 25 message credits (free plan) |
| v0 | 95/100 · tier pending v2 | about 15 min | 0 | $3.81 of a one-time $5 credit (free plan) |
What does that usage mean in money? See cost per working app: share of each free plan used and the dollar value at paid-plan rates, with usage screenshots.
One page per tool: Lovable review · Bolt review · Base44 review · v0 review.
| Tool | Test date | Methodology | Plan tier | Depth | Product / model version |
|---|---|---|---|---|---|
| Lovable | 16 Oct 2026 | Prompt-to-app v1 | Free | Tier B, 1 run | Hosted web app; the vendor shows no version number or underlying model, so none was recorded |
| Bolt | 16 Oct 2026 | Prompt-to-app v1 | Free | Tier B, 1 run | Hosted web app; the vendor shows no version number or underlying model, so none was recorded |
| Base44 | 16 Oct 2026 | Prompt-to-app v1 | Free | Tier B, 1 run | Hosted web app; the vendor shows no version number or underlying model, so none was recorded |
| v0 | 16 Oct 2026 | Prompt-to-app v1 | Free | Tier B, 1 run | Hosted web app; the vendor shows no version number or underlying model, so none was recorded |
Same prompt, same checks, a fresh account on each tool, desktop Chromium plus 375×812 mobile emulation, no paid top-ups. These tools update often. Results describe the tools on 16 Oct 2026 and may not match today's versions. We rerun when a tool ships a major update.
Evidence labels: scores, times, costs and bugs on this page are Observed: we saw them in our own runs, and the screenshots are below. Anything a vendor says about itself is marked Reported. What the labels mean. Disclosures: no tool on this page paid us, and none of these links are affiliate links yet; commission size never changes a score or tier.
All four passed every feature check and every data-isolation check. The difference was friction: Base44 and Lovable were live in under 4 minutes, Bolt in about 6, and v0 took about 15 with several manual setup steps. Each one had a different small issue, listed below.
Not sure which fits you? Take the 5-question quiz (based only on this test).
Lovable vs Bolt · Lovable vs Base44 · Lovable vs v0 · Bolt vs v0 · Bolt vs Base44 · Base44 vs v0 · Best for non-coders
In our first test (v1, 16 Oct 2026) Lovable, Bolt and Base44 all scored 100/100 and v0 scored 95/100. All four built a working, secure multi-user app from one prompt on the free plan with zero fix prompts. v0 was slower (about 15 minutes) and needed manual setup steps.
Base44 (about 3.5 minutes from prompt to live app), then Lovable (under 4 minutes), Bolt (about 6 minutes) and v0 (about 15 minutes, with a manual Publish).
Yes. All four finished our test on their free plans with no payment or top-up. Cost used: Lovable 2.10 credits, Bolt about 101K tokens, Base44 1 message credit, v0 $3.81 of a one-time $5 credit.
No. In all four apps, a second user could not see or change the first user's to-dos, including when we replayed requests with the second user's session.
See the app-builder tier list Cost per app Get your tool tested