Sprout, the “Tinder for jobs” built by a non-engineer in college — betting on distribution over building to reach $250K MRR in 8 months

Background & product

Sprout calls itself the ‘Tinder for jobs.’ You swipe through millions of roles; for the ones you like, Sprout matches your profile against the posting, asks a few focused questions to fill in gaps, and generates a resume and cover letter tailored to that specific job. Once you approve (or skip review), an AI agent fills out and submits the application on the employer’s own site. Everything that usually scatters across tabs is collected into one dashboard.

It was built by Nicole Cheung — a self-described ‘4x founder’ who launched several consumer AI apps with no formal engineering training while a student at Claremont McKenna College (Class of 2025, a double major in psychology and economics). Her first hit was the beauty app Glam Up: upload a selfie and AI analyzes your face and color season, then returns makeup advice and product picks. Her co-founder there was Aaron Paul, a Y Combinator–affiliated engineer. Within six months of its April 2024 launch — with zero ad spend and no outside funding — Glam Up passed one million users, hit #3 on the App Store, and reached $150K MRR.

Sprout (formerly Prep AI) ports that winning formula into a new category. Cheung’s thesis is blunt: now that ‘anyone can build an app in a weekend,’ the real constraint has shifted from shipping features to acquiring an audience. So with Sprout she treated distribution itself as the thing to engineer, as much as the product. Pricing comes in three tiers — Basic, Pro, Ultra — capped by applications per month, with monthly billing roughly 25% cheaper than weekly, and a subscription required to use it for real.

About eight months after launch, Sprout reached $250K MRR. Together with Glam Up, two apps were generating on the order of $400K a month combined — built without ads or venture capital.

The repeatable playbook

1. 1Design distribution before building — treat audience, not features, as the binding constraint
2. 2Pick a category that combines a universal self-improvement need, short-form ‘watchability,’ and willingness to pay for the outcome
3. 3Run UGC creators through a four-step loop: source → train via an in-house course → improve via feedback → mass-produce via referrals and analytics
4. 4Codify the winning angles so graduates hit fast (aim for the bar: a majority go viral within two weeks of onboarding)
5. 5On the product, collapse the most draining task into one swipe (for jobs: kill re-entry with AI generation + agent auto-apply)
6. 6Tier pricing by usage with monthly/weekly options; show value first, then convert (psychologically well-timed monetization)
7. 7Port the distribution playbook from app one into a different app-two category so the moat accrues in the *system*, not the app

Deep dive

【Deep dive】Sprout matters less as a job app than as a distribution machine that turns virality from luck into process. Here is the system Nicole reused across two apps, taken apart.

■ Core idea: the constraint moved from ‘building’ to ‘distributing.’ Cheung’s consistent claim: in an era when anyone can build an app in a weekend, what decides the winner is not features but audience. So she engineers distribution with the same intensity as the product. The conclusion she earned with Glam Up — that with zero ads and no funding you can still reach a million users if you systematize distribution — she simply re-ran with Sprout. It’s the inverse of ‘build it well and they’ll come’: decide how you’ll spread it first, then build.

■ The four-step creator system (a factory for virality). The heart of Sprout’s growth is a machine that *trains* and runs 200+ UGC creators. The reported shape has four stages: (1) source the creators, (2) onboard them through a structured course — effectively an in-house school that teaches the winning angles, (3) manage each creator with feedback loops to improve their posts, and (4) optimize with referral incentives and analytics so volume and quality compound. The effect shows up in the numbers: more than 50% of creators who complete the onboarding course put out a viral video within their first two weeks. Sporadic virality became a repeatable supply line.

■ The resulting reach. When the factory runs, it hit a peak of 400 million views in a single week, and it still sustains 400–500 million monthly views across TikTok, Instagram, and YouTube. The key is that this is not one mega-hit but the sum of many small wins: more creators means more supply, and referrals pull in the next creators — so headcount itself becomes the next wave of acquisition, a positive loop.

■ Product-side friction removal: collapsing an application into one swipe. Distribution alone doesn’t keep people paying. On the product side, Sprout ruthlessly stripped out application friction. The per-posting Workday gauntlet, the per-role resume and cover-letter rewrites — that whole chain is compressed into three moves: swipe, AI generation, agent auto-fill. Erasing the most draining part of job hunting — re-entering the same information — is what justifies a recurring subscription.

■ Pricing and monetization design. Three tiers — Basic, Pro, Ultra — capped by applications per month. Monthly billing is ~25% cheaper than weekly, nudging committed users toward monthly. And because real use requires a subscription (you can finish onboarding, but you must pay to actually apply at volume), the funnel shows value first, then charges. The behavioral-psychology and ‘emotionally well-timed monetization’ that shaped Glam Up runs underneath this too.

■ Why the jobs category (the cleverness of genre choice). Beauty (Glam Up) to jobs (Sprout) looks like a leap, but the aim is consistent. Both have (a) a near-universal self-improvement need, (b) short-form ‘I tried it’ content that pops, and (c) strong willingness to pay for the outcome. To distribute via a creator factory you want a theme that’s easy for UGC to mass-produce and fast for viewers to see themselves in. Job hunting fits exactly — and pairs naturally with student creators.

■ The takeaway for indie builders. Sprout’s lesson is ‘a distribution moat over a technology moat.’ Rather than training a frontier model yourself, translate existing AI into a frictionless experience (minimizing the build-side differentiation) and instead invest in a repeatable distribution asset: a trained, self-replenishing creator engine. The distribution playbook built for app one worked unchanged in app two’s category — meaning the moat had accrued not in ‘the app’ but in ‘the system for distribution.’