AppAlchemy: A Mobile App Builder Built in 2 Weeks Hit $10,000+ MRR in a Year — Finding an Open Seat

Dollar amounts are paired with rough yen equivalents at 1 USD = 150 JPY.

What launched in 2 weeks turned into over $10,000 MRR in a year

AppAlchemy is a tool for building native mobile apps entirely in the browser — no Xcode, no Mac, no dev environment setup required. It’s built by a single developer, Diego Roshardt. MVP build time: about two weeks. As of July 2026, roughly a year after launch, monthly revenue has reached over $10,000 MRR (about ¥1.5M). User count is “a few thousand,” in his own words.

The granularity of the numbers here is coarse. Exact customer counts, per-plan pricing, and month-by-month MRR progression are all undisclosed. Still, this case is worth covering because its winning formula is explained not by “what was built” but by “where it was placed.”

What’s confirmed

The technical stack migrated from React to Next.js, and the backend moved from Firebase to Supabase. The reason for the latter is straightforward: Supabase paired better with AI coding tools. Development itself uses Claude Code and Cursor.

Spotting the open seat

Roshardt is self-taught. He learned to code on YouTube, shipped several products in college, most of which failed, and after graduating, cold-emailed startup founders with his portfolio and landed a job as a software engineer. But within a few months he felt “unfulfilled” and quit.

Around that time, he was watching the boom in web-based AI app builders. Tools like Lovable were drawing major attention, but no equivalent product existed for mobile apps. Right next to the same demand curve, there was still an empty seat nobody had taken. That’s the origin point of AppAlchemy.

He left his hometown, rented a studio apartment in Austin, and committed to the business full time. “I only had a few months of runway, so it had to work” is how he describes that period.

Two moments that shifted the tide

There are two turning points in this business, one a deliberate design decision, the other coming from outside.

Charging from day one. AppAlchemy has no free trial period. From launch, it offered a monthly subscription and a discounted annual plan, going paid immediately. Roshardt’s reasoning: “paying customers are the ultimate proof that a product has demand.” Early revenue was indeed a thin trickle, “a handful of sporadic paying users.” Stacking up free users would have made the numbers look bigger, but he chose not to.

Getting picked up by a major account on X. At one point, an influential X account posted about AppAlchemy, and a large wave of users arrived overnight. This is the visible break point. To be candid, though, the source article doesn’t give exact user-count or MRR figures around this moment. There’s also a discrepancy between his own account of the timing and the dates shown on the tweet referenced in the article, so it can’t be pinned down to a specific day. What’s confirmed is only the sequence: “early sporadic paying users turned into something resembling an actual business after this influx.”

These two aren’t independent, their order matters. Because the billing mechanism was already in place, the overnight traffic converted directly into revenue. Had it been distributed free, that same traffic likely would have ended in nothing more than “the user-count graph spiked.”

Breaking down why it grew

Laid out at the surface, the tactics read as “posted on Reddit and X.” Go one layer deeper, and three structures appear.

One structure is borrowing a reference point. The category of web-based AI app builders already had completed market recognition. AppAlchemy could be explained in one line: “the mobile version of that.” It started with the cost of educating the market about a new category at close to zero, which is also part of why the MVP could ship in two weeks. Something that requires no explanation lets you strip features down.

Another: the friction being removed was concrete. The barrier to entry for mobile app development sits before “can you write code”, buying a Mac, installing Xcode, handling certificates. AppAlchemy folded all of that into a single browser tab. Because the friction removed was something anyone could instantly relate to, explaining the value takes little time.

The last is how he used Reddit. Roshardt’s method is straightforward: first, become an active Reddit user yourself. Working from the premise that “a post from a completely fresh account is almost certainly auto-filtered,” he uses a warmed-up account. Post content is product updates, videos of himself actually using it, and general discussion of vibe coding and AI tools. As he puts it, “provide value in your own niche, and casually mention the product along the way.” It’s simply about not leading with the pitch, but that’s precisely the condition that gets through both the filter and the reader.

What’s undisclosed, and the weak points

The biggest thing to watch in this case is how few verifiable numbers there are. Beyond the single figure of over $10,000 MRR, month-by-month progression, churn, and per-customer value are all undisclosed. So while the outcome (“$10,000 MRR in a year”) is confirmable, whether that breaks down as annual-prepay-heavy or steady monthly accumulation can’t be determined.

The structural risk is also not small. The AI app builder space is easy to enter, and multiple players already compete on the web side. The mobile-side vacancy that gave him an edge is exactly the kind of advantage that disappears once later entrants move in. On top of that, the product’s foundation sits on two external dependencies: AI models and App Store/Google Play review rules, neither of which is controllable.

Roshardt names his own missteps too. One is spending too much time on feature development rather than marketing. The other is the isolation of being a solo founder, with no community around him for a long stretch. His reason for recommending “build in public” is less about growth and more about staying connected to people, as he describes it.

What’s replicable, and what isn’t

What’s transplantable is the idea of searching for an open seat in an adjacent category. Rather than launching something entirely new, taking the “not-yet-filled axis” within a category where demand is already proven cuts both explanation cost and development cost. Web to mobile, or something popular in English-speaking markets moved into Japanese-speaking markets. There are multiple axes to take. Charging from day one is also directly replicable for anyone short on capital.

What’s harder to replicate is the timing and the traffic hit. Being present at the exact moment the AI app builder category was taking off, and getting picked up by a major account, aren’t things you can engineer intentionally. You could say sustained Reddit posting over time raised the odds, but the overnight traffic itself isn’t a condition you can build into a plan.

And it shouldn’t be forgotten: this is a story of quitting a job and relocating with only a few months of runway, committing fully. This decision is being told because it succeeded. Cases where someone made the same bet and ran out of money don’t get written up. When reading AppAlchemy’s numbers, that survivorship bias needs to be discounted.

Related reading

• PhotoAI $132K/month, zero employees, 37,000 commits a year. Dissecting Pieter Levels’ “run it solo” management style
• Zero marketing spend, an anonymous solo developer, $2M ARR. Carrd’s “don’t add features” strategy