How to Build an MVP: One SaaS From Prompt to Landing Page in a Day, Every Credit Counted

Every guide for this question is an eight-step list and a price range. This one is a single build: a multi-tenant SaaS called Cadence, started at 12:26 on one day with a library prompt, reviewed for SEO, scanned for security, made installable on a phone and given a landing page by 19:28, with the credits for each step read from the billing page and the screenshots on the page that covers it. It ends where honest MVP stories end: with a working product and no users yet, and the step that comes next.

Updated 26 September 2026

How to build an MVP hero: the title beside a ledger card, 12:26 foundation with auth, roles and invites at 10.4 credits, 14:29 SEO review 5 of 6 passing at 1.1, 18:06 security chat 1.3 with the Deep scan free, 18:36 installable PWA 3.1, 19:19 landing page six sections 4.5, what is still missing before a first user

What is an MVP for, in the words of the person who named it?

Eric Ries, 3 August 2009: “the minimum viable product is that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.” And the sentence most guides forget: “MVP, despite the name, is not about creating minimal products.” The build below is judged by that standard. Not whether it is small, but whether it can teach something about customers for the least effort, and effort here is counted in credits.

Why venture-backed companies shut down, 385 post-mortems since 2023, percent citing each reason

Why venture-backed companies shut down, 385 post-mortems since 2023, percent citing each reason
Category	%
Ran out of capital	70
Poor product-market fit	43
Bad timing	29
Unsustainable unit economics	19

CB Insights, read 26 September 2026: 431 companies analysed, 385 with identifiable reasons; companies cite more than one reason, so the bars do not sum to 100.

Product-market fit and timing are learning failures: the company built before it knew. The MVP is the instrument for knowing earlier, and the cheaper the instrument, the earlier you can afford to use it.

What was built, in what order, and what did each step cost?

One project in Marco's personal Lovable workspace, on 26 September 2026. Each row is a prompt sent to the builder, the credits it cost as read from the billing page, and what came back. Each links to the page that shows the run.

The day, in order

Times from the Lovable chat; credits from the billing page.

The MVP ledger, 26 September 2026: 20.4 credits in total
	Prompt	Credits	What came back
12:26 Foundation	SaaS MVP with User Auth, from the library See the run	10.4	Asked 4 questions, enabled Cloud, self-tested sign-up and invites, roles checked in the database
14:29 SEO review	Run an SEO and AI search review of this app and show me the findings. Do not change any code yet. See the run	1.1	5 of 6 checks passing; the one failure was the sitemap, deferred until publish
18:06 Security	Run a deep security scan of this app and list every finding by severity. Do not change any code. See the run	1.3	The chat ran its own checks (clean) and said the Deep scan must start from the Security tab; run there, free, 3 minutes, no findings; the dependency panel listed 2 known issues on one package
18:36 Installable	Make this app installable on a phone as a Progressive Web App ... Do not change any feature, table or policy. See the run	3.1	Manifest, icons, service worker gated to the published app, Add to Home Screen button; verified in the preview
19:19 Landing page	The six-section SaaS landing page prompt, in a separate project See the run	4.5	Five pre-build questions, six minutes to a preview; server-rendered, FAQ schema, one CTA type; 394 words

Credits per step

Credits per step
Category	credits
Foundation	10.4 (12:26)
SEO review	1.1 (14:29)
Security	1.3 (18:06)
Installable	3.1 (18:36)
Landing page	4.5 (19:19)

Read from the billing page after each step on 26 September 2026. The Deep security scan itself was free; the 1.3 is the chat message that asked for it.

The foundation is half the bill and most of the product. Everything after it is the part agency guides put in a later phase: review, harden, install, market. On a builder that asks questions first, each was one message and under five credits.

What does the result look like?

The Cadence marketing landing page after the last step: headline, one trial button, placeholder proof line, problem section — After the last step: the landing page a first user would arrive on, 26 September 2026.

Between the two: the SEO review passed five of six checks and deferred the sitemap until publish; the Deep security scan found nothing at any severity while the dependency panel listed two known issues on one package; the app gained a manifest, icons and an install button. The security and mobile pages have those screenshots.

What is still missing?

A first user, and the things a first user exposes. The product works for the person who built it. Nobody else has signed up, because nobody has been sent to the landing page; it has no traffic and therefore no conversion rate. The proof line on that page still says it is a placeholder, because there is no customer to quote. Payments are not wired; the Lovable and Stripe page is that step when it is earned. None of this is a gap in the build. It is the point at which building stops being the bottleneck.

What comes after the ledger

The first 100 users and product-channel fit guides, condensed.

The first 100 users page is the long form; product-channel fit is why the channel is chosen before the next feature.

What do the top results for this term leave out?

Cleveroad, How to Build a Minimum Viable Product in 2026. About 4,500 words, updated 6 January 2026, eight steps, two to six months, $50,000 to $200,000; portfolio projects as evidence; the CTA is the agency.
UXPin, How to Build an MVP. About 3,400 words, 17 June 2026, eight steps, $5,000 to $150,000 with no sources for the ranges; the CTA is the tool.
Uptech, How to Build an MVP. About 6,500 words, updated 9 March 2026, eight steps, $5,000 to $100,000 and three to twenty-four months from the agency's own experience; its own products as examples; the CTA is the agency.

None of the three shows a build the reader can repeat with the cost attached, and none ends with the product unused, which is where every MVP actually is on the day it is finished.

What we did not publish

A conversion rate, a user count or revenue. The product has none yet, and the page says where that changes.
A timeline in hours. The ledger has clock times because the chat records them; the hours of attention between steps were spent writing the pages that document them, and are not separable.
A comparison build on another tool. The same day's work on Bolt, v0 or Replit would be the fair test, and it did not happen.

Sources

All read on 26 September 2026. The ledger is composed in src/data/mvpLedger.ts from the run files each step already has.

Frequently asked

What is an MVP?↓

Eric Ries, 3 August 2009: "the minimum viable product is that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort." And, from the same post: "MVP, despite the name, is not about creating minimal products." The build on this page is an MVP in that sense: a working SaaS foundation with a landing page, built to learn whether anyone signs up, not to be small.

How long does it take to build an MVP?↓

One working day on this page: the foundation at 12:26, the landing page previewed by 19:28, with an SEO review, a security scan and an installable build between. Agency guides on the same query say two to six months; the difference is that they are pricing a custom build with a team, and this is one person with a library of prompts and a builder that asks questions first.

How much does an MVP cost?↓

20.4 Lovable credits for everything on this page, read from the billing page after each step, on a $25 Pro plan that includes 100. The agency guides ranking for this query quote $5,000 to $200,000; the cheapest of their ranges is for no-code, and this page is what that range looks like when it is measured rather than estimated.

Why do MVPs fail?↓

CB Insights' analysis of 431 venture-backed companies that shut down since 2023, 385 with identifiable reasons, read 26 September 2026: ran out of capital 70%, poor product-market fit 43%, bad timing 29%, unsustainable unit economics 19%. Two of the four are about not learning fast enough, which is the thing an MVP exists to do.

What is still missing after this build?↓

A first user. The foundation works, the landing page exists, the security scan is clean and the app installs on a phone, and nobody has signed up because nobody has been sent to it. The next step is a channel, not a feature: the first 100 users page is where this story continues, and the product-channel fit page is why.

Can I follow the same path?↓

Yes, with the same prompts: each row of the ledger links to the page with the prompt, the screenshots and the credits. Run the foundation, ask for the SEO review, run the security scan from the Security tab, add the PWA if your users are on phones, then build the landing page in a separate project so its analytics are clean.