CJ posted a single tweet describing his "ideal Cursor workflow" — essentially a screen-recording-style demo showing how he solved AI hallucinations using structured documentation. The tweet was not a product pitch; it was a tutorial framed around the problem. It received ~400 retweets and generated ~300 immediate signups. He had no product at that point. Within 30 minutes he used Bolt to spin up a basic landing page and collect waitlist emails. Over the next two weeks he accumulated 1,800 waitlist signups before writing a single line of product code. He then DM'd and engaged every person who commented, asking whether they had a better solution to AI hallucinations. No one did — which served as his second-stage validation. The first 100 paying users on launch day (December 24, 2024) came directly from that waitlist.
CodeGuide
AI-generated knowledge base docs to reduce hallucinations in AI coding tools
8 moves, in order
- Pre launch / Idea ValidationTwitter — organic
Posted a single tweet describing his ideal Cursor AI workflow and the AI hallucination problem he faced. Used a screen-recording/demo format. No product existed — just shared the problem and his manual solution.
~400 retweets, ~300 immediate signups - Pre launch / WaitlistLanding page waitlist
Built a landing page in ~30 minutes using Bolt (no product code written) and funneled tweet traffic to it to collect waitlist emails.
1,800 waitlist signups in 2 weeksUsers 1.8k users - Pre launch / ValidationDirect dm outreach
Manually DM'd every person who commented on the tweet asking if they had a better solution to AI hallucinations. Used their responses to confirm no superior alternative existed and to refine the product direction.
Confirmed product-market fit; no competitor solving the problem better - Launch DayEmail waitlist
Launched the product on December 24, 2024 and notified the waitlist. Waitlist subscribers converted because they already trusted CJ and identified with the pain point.
100 users on day 1Users 100 users - Post launch Growth (Days 1–49)Twitter — threads
Implemented 'tutorial marketing': 4 threads + 3 long-form posts per week on X. Each piece follows a hook (problem) → blueprint (ideal solution) → product positioning formula. Designed visual assets in Canva/Figma. Drafted content one day, edited the next.
$17,000 MRR in 49 daysMRR $17k - Post launch Growth (Day 49–90)Twitter — threads
Continued the same tutorial marketing flywheel — doubling down on content formats and topics that performed well (bookmarkable, shareable tutorials). All traffic to site came exclusively from X. No paid marketing spend.
$42,000 MRR at ~90 days, 4,800 community membersMRR $42k Users 4.8k users - Ongoing Retention & MonetizationPartnership integrations
Secured official partnerships with Cursor AI, Windsurf, Bolt, Lovable, and other top AI coding tools, bundling perks into the annual plan. Used this to justify a 40% discount on yearly memberships and increase perceived value.
Low churn rate cited; supports annual plan conversion - OngoingEmail newsletter
Used ConvertKit to email the full waitlist/user base. Treated email as a retention and re-engagement channel for users who signed up via the landing page.
Supports conversion from free signups to paid; specific numbers not stated
Active X/Twitter presence in the AI coding niche — had already been publicly posting about AI prompting since 2022 and had built credibility in that community before CodeGuide existed. Also had hands-on experience running an MVP agency, giving him deep firsthand insight into the exact pain point he was solving.
twitter_threads
No paid marketing used at all (explicitly stated, not a failure per se, but no paid channels were tested). 10 out of 11 prior startups failed before CodeGuide. No specific failed channels for CodeGuide itself were mentioned.