Solo founder, bootstrapped to $80M exit in six months: Maor Shlomo on building Base44

Executive overview

Most vibe coding tools rely on third-party integrations for backend functionality. Base44 built everything full-stack — database, user management, analytics — making it more capable for complex real-world apps.

Maor Shlomo built Base44 alone, without raising a single dollar, reached profitability within weeks, and sold to Wix for $80M+ six months after starting.

The core insight: a small team managing AI can move faster than a well-funded competitor, and bootstrapping removes the pressure that kills creative momentum.

How the idea emerged

• Two real problems triggered the idea: building a web app for his girlfriend's small business and watching a scouts organisation pay agencies $1M for tools that LLMs could generate
• Previous company (Explorium) raised $130M; Base44 was a deliberate reset toward building things he loved
• Spent a year in IDF reserves post-Oct 7, then travelled — returned wanting to code again
• Knew the vibe coding category was crowded but saw a differentiated angle: batteries-included, full-stack architecture vs. competitors' frontend-plus-Supabase approach

Getting the first users

• Started with three close friends, sitting with them every other day while they tried to build things — fixed bugs live, built features on the spot
• Key signal before investing in marketing: users spontaneously sharing the product with others
• First paid Product Hunt launch netted ~15 users; second launch was flagged as bots by the algorithm (community had driven ~500 more upvotes than second place), resolved after community contacted PH support
• Tried influencer post (~$2K) and paid ads — neither worked; stopped quickly

Building in public as a growth engine

• Started sharing the Base44 journey on LinkedIn at around February; growth accelerated immediately after
• Honesty over polish: shared the good, the bad, and the ugly — no VC-funded metrics theatre
• Audience were fellow builders, so the content matched perfectly; this alignment is not universal
• Picked LinkedIn as the single channel and doubled down — cross-posting to Twitter had negligible ROI
• Starting network was a few thousand LinkedIn connections, not zero

Virality and community tactics

• Offered free credits to users who posted publicly about what they built (didn't need to mention Base44)
• Early community ran through WhatsApp — fast feedback loop, good uptime monitoring by proxy
• Scaled later to Discord and Reddit as numbers grew
• Velocity itself became a marketing asset: shipping features daily kept people publicly speculating and sharing

The "for good" hackathon

• Ran a hackathon focused on social-impact apps when at ~5–10K users, with a $5K prize from his own profits
• 3,000 teams registered; Amazon, Google, MongoDB, and Deloitte sponsored and opened offices globally
• Produced standout apps (e.g., an Alzheimer's memory game built for a founder's grandmother)
• Strong earned media and community attachment; described as one of the top career moments

Solo founder realities

• First hire came 1.5 months before acquisition — a technical product generalist (not an engineer)
• Biggest solo challenges: no one to share stress with, no on-call support, brutal prioritisation between what he wanted to work on (product) vs. what he had to (marketing)
• Daily ritual: explicitly ask "what do I need to work on?" vs. "what do I want to work on?"
• Used RescueTime to block Twitter/LinkedIn during deep work; severe ADHD made focus infrastructure non-optional
• Built custom Base44 apps for his own workflows — content drafting, tone-matching past posts, LinkedIn-to-Twitter reformatting

Tech stack

• Render.com for all infrastructure (platform, user apps, website — kept isolated)
• MongoDB for flexible schema handling; important when LLMs keep changing data structure
• Cursor for coding; spent 20–30% of time structuring the repo to be LLM-friendly
• Python backend; JSX (not TypeScript) frontend — models write JSX more reliably
• Mono-repo approach: keeping frontend and backend together gives the AI better context
• Built a prompt-routing layer: Claude for initial generation and UI; Gemini for complex logic and bug loops; smaller fast models (Flash, GPT-4o Mini) to patch code inside files
• Wrote as little code as possible per feature — opinionated infrastructure handles CRUD, auth, DB so the LLM only writes the delta

Activation insight

• Originally showed users a generated PRD/user-flow step before building — improved app quality but hurt conversion
• Ditched it: the aha moment in vibe coding is the instant surprise of seeing a working app; any intermediate step dilutes that
• Rule: get users to the aha moment in under 2–3 minutes, even at some cost to initial app quality

Bootstrapping vs. raising

• Reached ~$200K monthly profit by May (month ~4); was viable as a standalone business
• Bootstrap is not right for every product — B2B/enterprise sales requires capital; viral consumer products are a better fit
• Being default-alive removes the psychological weight of investor money; easier to sustain energy over time
• Solo + AI: hasn't written a single line of HTML or JavaScript in three months; funding gap vs. competitors is increasingly irrelevant

The acquisition

• Wix reached out after community members publicly suggested the fit (shared customer base, Israeli ecosystem)
• Chemistry with CEO Avishai Cohen was central — multiple informal dinners before any deal was on the table
• Negotiating position: genuinely fine with either outcome — removes desperation, improves terms
• Deal structure: published ~$18M upfront + earnout (air note) tied to Base44's future performance — keeps Maor financially and personally invested in growth
• Signed papers the morning after Iran-Israel hostilities broke out (lawyers had refused to finalise at 2am)