Lewis launched Audio Pen as one of four or five tiny tools he built in a single week and hosted on his personal website. While building each one, he shared them live on Twitter as part of the "build in public" community. Audio Pen's MVP received far more engagement than the others, so he DM'd interested users to understand their use cases before building the full version. At the end of the 12-hour hackathon (~10 hours in), he spun up a waitlist so early followers could sign up before the full tool was ready. Before the 12 hours were up, he began receiving unsolicited Stripe payment notifications from beta testers who had been given free access — they paid voluntarily. He attributes this early conversion to product-market fit, his established credibility from months/years of building in public on Twitter, pre-launch hype he had generated, and social proof from the Half-Day Build community who were visibly cheering him on.
Audio Pen
Voice-to-text AI tool that turns spoken thoughts into clean, styled writing
8 moves, in order
- Pre launch / Idea ValidationBuild in public — Twitter
Participated in the Half-Day Build online hackathon community on Twitter, joined a Slack/Discord with other indie builders, and committed to going from idea to revenue in 12 hours starting at noon on a Sunday.
Created accountability, community support, and a live audience for the launch - MVP Build DayPersonal website
Built four or five small tools in one week and hosted them all on his personal website homepage, sharing each one on Twitter as he built them. Audio Pen was one of these tools.
Audio Pen received significantly more engagement and positive signal than the other tools - MVP LaunchTwitter posts
Posted about Audio Pen's MVP on Twitter while building it during the hackathon, leveraging his existing build-in-public following and the Half-Day Build community for immediate social proof and visibility.
Unexpected traction — people responded positively and validated the product concept - Pre launch WaitlistEmail waitlist
About 10 hours into the hackathon, launched a waitlist so interested users could sign up before the full product was ready. Did not ask for payment — just asked people to test the product.
Collected early beta testers; received unsolicited Stripe payment notifications before midnight - Early User ResearchTwitter dms
After seeing signal from the MVP post, Lewis DM'd interested users on Twitter to ask why they liked it and what their use cases were. Used this to inform the full product design.
Identified core use cases (idea capture on the go, meeting notes, writer's block) that shaped the product roadmap - Full Product LaunchBuild in public — Twitter
Designed the full app in Figma (using Pinterest for design inspiration), built it using Bubble (no-code), and launched publicly with a paid tier. Continued sharing the journey on Twitter as an indie builder, not a faceless company.
Got 20th paying customer — identified as the inflection point where Lewis knew it was working - Growth / ScalingFreemium conversion
Maintained a free tier giving users a limited taste of the product (shorter recording time, no advanced styles), with a paid plan at $99/year or $159/2 years. Gradually raised price as the product was refined.
Grew to 200,000 total users with 5,000+ paying customers at ~$15,000/monthMRR $15k Users 200k users - Retention / DifferentiationProduct focus
Deliberately resisted expanding into adjacent markets despite temptation. Focused on doing one thing — voice-to-structured-text — exceptionally well. Maintained indie builder transparency and personal brand as a differentiator.
Users cite the personal builder relationship as a reason they pay; sustained $15K/month as a solo, part-time operatorMRR $15k
Lewis had an established Twitter presence from years of building in public, giving him a warm, engaged audience that trusted him before Audio Pen launched. He also had a ready-made community through the Half-Day Build hackathon group on Slack/Discord, providing immediate social proof and word-of-mouth at launch.
twitter_build_in_public
Lewis explicitly notes he was tempted to expand into adjacent markets (seeing other apps doing well nearby) but consciously avoided it — implying product scope creep was a failure mode he resisted. He also failed with 15–20 prior products before Audio Pen, though specific channels for those failures are not detailed.
I shipped my app in 12 hours and now it makes $15K/month
Jenny AI
AI-powered writing and research assistant tools for students and researchers
Chatbase
AI chatbot builder — train a ChatGPT on your own data/documents
StealthGPT
AI tool that makes AI-generated text undetectable by AI detectors