Steal this idea: Meet Oscar's $5 queue-jump deposit

Meet Oscar is an AI email assistant that drafts replies in your voice, using the thread and documents you give it, such as price lists and FAQs.^[2] Its founder, Giovanni Brees, says he has zero coding experience and built the first version over a weekend in Lovable.^[1] He says it reached $45,000 in monthly recurring revenue within 60 days of launch.^[1]

The weekend build is the part people share. The part worth stealing happened before it: he didn't build anything until strangers had paid him $5 for a better spot in line. This page breaks down that test, two other AI-built products that sold before they were finished, and a step-by-step version you can run with a landing page and a Stripe link.

The pattern in one sentence

Put up a waitlist with visible positions, then offer a small payment to move to the front, and treat the number of people who pay (not the number who sign up) as your real demand signal.

How Meet Oscar ran it

There was no app yet, just landing pages and a waiting list.^[2] Here is the flow as he describes it on his site and in a SaaS Mastery interview:

1. Landing pageone simple page about the pain: email overload^[1]
2. Waitlistsigning up put you at the back of the line, e.g. number 1,000^[1]
3. $5 depositpay to move into the first 100 users, first in when it goes live^[1][2]
4. Referralsa referral link; five sign-ups earned extra free months^[2]

He calls the sign-up the first validation and the payment the second.^[2] The referral link was a third: in the interview he says people who brought in five sign-ups got something back, for example two extra months of access, and each referred person was then offered the same $5 deal, so the loop started again.^[2]

The part most people skip: 15 audiences, one question

The deposit only tells you something if you can compare it across audiences. He used Lovable to make more than 200 variations of the landing page for 15 distinct customer types, from lawyers and insurance agents to support teams and ordinary Gmail users, each with its own wording and images.^[1] In the interview he says the exact number would need checking, "let's say main groups about 15", with subgroups such as solo lawyers, lawyers in firms and insurance lawyers.^[2] The method was simple: one main page, then asking Lovable for a variation with a new image and copy.^[2]

The metric that decided things was prepayments per audience, not clicks.^[1] That's how they found out they were wrong. They expected support agents to be the biggest market; the lead time to sign them up was too long. Consumer Gmail users, whom they never expected to work, turned out to be "desperate for email help".^[1][2]

Two details from the interview are easy to miss. Groups with a low deposit rate got follow-up emails and a second offer, and some caught up; he says European audiences were more careful with card details than US ones and needed more emails.^[2] And the deposits did more than validate: he says the small payments paid back part of the pre-launch ad spend.^[2]

He used the same trick again after launch

Once Meet Oscar was live, customers started asking for features they said they couldn't live without. His answer was the queue-jump again, in a bigger size: pay for a year up front instead of monthly, and your request moves up the priority list.^[2] The point, as he tells it, was to see which requests people would commit money to, and to weigh 100 people asking for the same thing above one or two people "yelling really hard".^[2] He also kept the first version tiny: WhatsApp and Instagram ideas were dropped, and the MVP did one thing, email replies in your own voice.^[2]

Two other AI-built products that sold first

Klar, a study agent from three university students in Uppsala who call themselves non-technical, took real payments (Apple Pay, card) from students at a university career fair before the product existed. The pitch: cheaper than the ChatGPT Plus subscription students already paid for. After launch they let people in in batches to make it feel exclusive.^[3] That's Shipped & Paid's summary of the founders on the Foundore podcast; in August CEO Isabel Storgårds posted that Klar had crossed $500K ARR (founder-reported).^[4]

Backchannel, a guided job search, got its first customer from a Loom video. Mindaugas Petrutis, a former recruiter with no coding background, shared a recording of his Lovable prototype, and the first person who saw it asked to pay.^[5]

Three different shapes, same idea: ask for money while the thing is still cheap to change. Meet Oscar did it at scale with ads, Klar did it face to face, and Backchannel did it with one video and one link.

How to run it yourself this week

Steps 1, 2 and 6 follow what Brees describes; the Stripe mechanics and steps 4, 5 and 7 are our suggestions for doing it without code.

1. 1

   Write one plain page about the pain

   One page, no time on the logo (his first one took five minutes in PowerPoint), the problem in the headline and a waitlist form.^[1]
2. 2

   Clone it for 3 to 5 specific audiences

   He suggests starting with about five pages on a low budget, each aimed at the most likely buyer, as narrow as you can go before ad costs climb.^[2] Change the picture and the words, keep the offer the same, so the deposit rate is comparable.
3. 3

   Show a queue number on the thank-you page

   In his version a sign-up was placed at the back of the line (his example: number 1,000).^[1] Make the offer right there: pay $5 and you're in the first group.
4. 4

   Take the $5 with a Stripe Payment Link

   Payment Links need no code and come with receipts and refunds from the Dashboard.^[6] Add the waitlist ID as client_reference_id in the link so each payment maps back to a person, and UTM tags so you know which page and audience it came from.^[8] If the front of the line really is 100 seats, Stripe can deactivate the link after a set number of payments.^[7]
5. 5

   Say exactly what the $5 buys

   Earlier access, and whether it's credited against the first month or refunded if you never ship. Our suggestion; Brees doesn't say how Meet Oscar handled it, and we didn't find it on his pages.
6. 6

   Count deposits per audience, then build for the winner

   The number to watch is prepayments per audience.^[1] Give slow groups one or two follow-up emails before ruling them out.^[2]
7. 7

   Build only the one thing they paid for

   For Oscar that was reply quality, then security.^[2] Let the first 100 in, then add a referral reward and keep offering the deposit to new sign-ups.^[2]

// Payment Link URL: https://buy.stripe.com/<link>?client_reference_id=<waitlistId>&utm_campaign=lawyers
import Stripe from "stripe";
const stripe = new Stripe(process.env.STRIPE_SECRET_KEY!);

export async function POST(req: Request) {
  const event = stripe.webhooks.constructEvent(
    await req.text(),
    req.headers.get("stripe-signature")!,
    process.env.STRIPE_WEBHOOK_SECRET!,
  );
  if (event.type === "checkout.session.completed") {
    const s = event.data.object;
    // s.client_reference_id is the waitlist id you put in the link
    await moveToFrontOfQueue(s.client_reference_id, s.amount_total);
  }
  return new Response("ok");
}

You don't need the webhook to start. With a handful of deposits you can check each link's payments in the Stripe Dashboard and move people by hand.^[7]

Gotchas before you copy it

• Don't fake the product. In the interview he describes follow-up emails to slow groups that talked about how he used Meet Oscar in his own company, before there was a commercial version, adding that "the user doesn't know".^[2] We wouldn't copy that part. When you take money before the product exists, say so plainly; a deposit from someone who knows it's a waitlist is the honest version of the signal.
• A vibe-coded MVP can leak. He says the first version was wired straight to the ChatGPT API and the API key "was pretty easy to hack", and that they had a security issue early on.^[2] Keep AI keys on a server, never in the page.
• Email products hit Google's review. Reading Gmail needs restricted scopes such as gmail.readonly or gmail.modify, and if you store or transmit that data on servers you must pass a security assessment.^[9] Google runs it on the CASA framework, with annual recertification.^[10] Brees says needing that approval was one reason he brought in a developer co-founder and rebuilt the product outside Lovable; the other was wanting their own model instead of depending on ChatGPT.^[2]
• $5 proves pain, not price. A coffee-sized deposit tells you people care; it doesn't tell you they'll pay your real monthly price. Our note: test the real price at launch (Oscar also tried VIP upsells then).^[2]
• He wasn't starting from zero. His site says he spent 15 years in growth and validation and helped bring more than 4,500 products to market, with over $504 million raised in crowdfunding.^[1] Running 200 ad-backed pages is his day job. A first-time founder should start with the five-page version.
• All growth went back into ads. After launch every dollar of revenue went into Meta and Google ads, and the founders took no salary.^[1][2] That's part of how the 60-day number happened, and not every founder can do it.

What to take from this

1. 01

   Interest is not demand

   A sign-up is a polite yes; $5 is a decision.^[1]
2. 02

   Test audiences, not features

   The winner was a group he didn't expect.^[1]
3. 03

   Make the line visible

   A queue number gives the deposit a reason.^[2]
4. 04

   Sell where buyers already are

   A career fair worked for Klar.^[3]
5. 05

   A link beats a pricing page

   Backchannel's first sale was one Stripe link.^[5]
6. 06

   Reuse the mechanic

   Pay yearly, move your feature up the list.^[2]