The visible loop is people sharing what they made. The hidden prerequisite is output quality. Attribution helps only after the first build creates pride. GPT Engineer also meant Lovable entered launch with a problem-aware audience and a public record of technical credibility, not merely a large following.
COMPANY PLAYBOOK 02 / 15
THE LOVABLE SYSTEM
Lovable
$200M ARR one year after commercial launch
The launch began eighteen months before the product existed. Every user output then became a distribution artifact.
THE QUICK READ
Lovable did not arrive cold. Anton Osika’s open-source GPT Engineer project established the problem, attracted builders, and created trust before the commercial product launched in November 2024. One year later, Lovable reported 100,000 new projects a day and 5M daily visits to Lovable-built products.
AudienceProductCommunity
01 / WHAT THE RECEIPTS SAY
02 / BACKLOG-READY PLAYS
2 plays worth stealing.
Open a play for the mechanism, the exact receipt, the failure mode, and a deployment brief Elena can put into your backlog.
07 Open-source the belief before selling the product
THE MOVE
Osika released GPT Engineer in June 2023, letting the market experience the text-to-software thesis while the eventual commercial product was still forming.
WHY IT COMPOUNDS
The open-source project reduced category education cost, supplied a community of high-intent early adopters, and made the commercial launch feel like the next chapter rather than an unknown product.
THE RECEIPT
GPT Engineer accumulated roughly 52K GitHub stars before Lovable’s launch; Lovable reported $4M ARR in four weeks and $10M within 60 days with a 15-person team.
STEAL THIS
Release the smallest useful expression of your thesis: a tool, benchmark, dataset, or protocol. Grow the problem community before asking it to buy the polished workflow.
DO NOT COPY BLINDLY
Open source is not free demand. It works when the free artifact and paid product advance the same belief but serve different jobs.
RECEIPT STATUS
Operator-reported. Anton Osika connects GPT Engineer, the launch audience, and Lovable’s early revenue trajectory.
Open the exact source ↗ ELENA DEPLOYMENT BRIEF
Turn this observation into a real experiment.
- Run it when
- The category belief can be expressed as a useful artifact before the full product exists.
- Owner
- Technical founder and developer community lead
- First sprint
- Release one narrow tool, publish the thesis, and interview the first 25 repeat users about the paid workflow they still need.
- Leading signal
- Repeat use, meaningful contributions, thesis citations, and paid-product waitlist conversion.
- Stop rule
- Stop investing if the artifact attracts a different audience or job than the commercial product for two cohorts.
08 Turn customer output into the top of funnel
THE MOVE
Lovable users publish real websites and apps, then share the thing they made instead of a screenshot of the software used to make it.
WHY IT COMPOUNDS
The output carries social proof, sparks “how did you build that?” curiosity, and recruits new builders. Distribution is attached to the customer’s pride rather than a referral prompt.
THE RECEIPT
At the one-year mark Lovable reported 100,000 projects created daily and 5M daily visits to Lovable-built sites and apps.
STEAL THIS
Make the finished outcome public, attributable, and easy to remix. Ask what users naturally want to show when they succeed.
DO NOT COPY BLINDLY
Bad outputs create negative distribution. The share loop only compounds after the first result is worth showing.
RECEIPT STATUS
Unfairly synthesis. Project and visitor volume are company-reported. The output-to-acquisition mechanism is our synthesis.
Open the exact source ↗ ELENA DEPLOYMENT BRIEF
Turn this observation into a real experiment.
- Run it when
- Users produce a finished artifact they are proud to show outside the product.
- Owner
- Product growth and design
- First sprint
- Make one output public, attributable, and remixable. Test the share flow with ten successful users.
- Leading signal
- Share rate after success, visits per shared output, and referred-user activation.
- Stop rule
- Remove or redesign attribution if shared outputs create negative sentiment or referred users activate below organic baseline.
03 / MAKE IT YOURS
Do not copy Lovable. Adapt the system to your constraint.
Elena learns your product, customer, funnel, and current bets. Then she chooses the relevant pattern, scopes the first sprint, and watches the leading signal.