​Decoding the Vibe: What We Learned from 100 Vibe Coders (and What We Built Next) by Pierre Burgy, CEO Strapi

Pierre wrapped up the evening with a data-driven exploration of the **vibe coding landscape**.
After interviewing **50 users** and collecting **500 insights**, his team identified five key takeaways:

1. 🌍 The **European bias is real** — Lovable dominates, but Bolt and Replit are close behind.
2. 👥 Two main user profiles: *the weekend tinkerers* and *the rapid prototypers.*
3. ⚙️ **Cursor and Claude Code** are game-changers for debugging and depth.
4. 🧪 Most users are still in **prototyping mode** — vibe coding as *“Figma for code.”*
5. 🚀 **Vibe-coding agencies** are already forming to help founders ship faster.

He concluded with a teaser: the launch of **Strapi’s upcoming AI project (`****.ai beta`) — born from these learnings and focused on empowering the next generation of AI-native creators.

Transcript

2. 1. Vibe Coding Year in Review. From Karpathy to Real-World Agents by Niels Rolland, CEO Paatch 2. Becoming a Better Product Manager Through Vibe Coding by Jules Boiteux, Founder Vibe Coding Academy 3. Turning raw data into life-saving decisions using n8n and vibe-coding by Claire Champourlier, Technical Startup Advisor at Interrest 4. How Enterprises Build on Top of Existing Systems with AI by Augustin Sorret, founder of HelloLeo 5. Decoding the Vibe: What We Learned from 100 Vibe Coders (and What We Built Next) by Pierre Burgy, CEO Strapi Agenda
3. What we learned from interviewing 50 Vibe Coders Pierre Burgy
4. About me Pierre Burgy Co-Founder & CEO of Strapi @pierre_burgy pierre@strapi.io
5. Why? What is the future of website creation?
6. How it started Looking for Vibe Coders
7. How it’s going Looking for Vibe Coders
8. How it’s going Looking for Vibe Coders
9. Looking for Vibe Coders 4 Weeks 500 Insights 50 User Interviews
10. Product Managers 32% Solo founders / CEO 32% Designers 18% Sales/Marketers 11% Engineers/CTO 4% Others 4% #1 ROLE & BACKGROUND What’s your current role, and what do you typically work on day-to-day?
11. #2 TOOLS IN USE Which “vibe-coding” tools have you used recently, and what for? Lovable 52% Bolt 48% v0 39% Replit 13% Base44 6% Cursor 39% Codex 10% Claude Code 9% Windsurf 4% Vibe Coding AI Coding
12. #3 ADOPTION JOURNEY When and why did you start using these tools?
13. #4 USE CASES What kind of projects did you (or do you) build with them?
14. Frequency 81% Daily 29% #5 ENGAGEMENT & RETENTION How often do you come back to improve or revisit your projects? Why?
15. 🥇 First output from first prompt 🥈 History (rolling back to previous version) 🥉 Integrations (GitHub, Supabase, Stripe) Uploading images & drawings Visual edit (without burning credits) #6 DELIGHT FACTORS What do you love most about the tools you’ve used?
16. 😡 Being stuck 😡 Lack of precision (unexpected edits) 😡 Burning credits due to AI’s mess 😡 Lack of UI consistency 😡 Security concerns (code & integrations) Huge improvements in the last few months #7 GAPS & FRUSTRATIONS What’s missing or frustrating in these tools? Why does it matter
17. 💰 Most of these tools have a base price of $20-$25 ✅ Paying $20 is worth the value 🔄 Users often switch from one tool to another #8 PRICING & WILLINGNESS TO PAY How much are you paying (or have paid) for these tools? Was it worth it?
18. $20 + additional credits/tokens/usage #9 PRICING PREFERENCES If you could design your ideal pricing model for such a tool, what would it look like?
19. #10 FINAL THOUGHTS If you could wave a magic wand and instantly improve one thing about these tools, what would it be? 😡 Being stuck 😡 Lack of precision (unexpected edits) 😡 Burning credits due to AI’s mess 😡 Lack of UI consistency 😡 Security concerns (code & integrations)
20. Conclusion 1. The European bias is real We were convinced Lovable was way bigger than Bolt. Turns out… not so much. 2. Two types of users emerge Those waiting for free weekends and burning through credits. Those prototyping fast and shipping even faster. 3. Cursor is a game-changer Many users switch to Cursor (or Claude Code) when they get stuck, especially for unlocking tricky bits, fixing security issues, or moving beyond surface-level edits. 4. Production isn’t the goal (yet) Most users are still in prototyping mode — using vibe-coding like an interactive spec tool. Think Figma + code + customer feedback loops. 5. Vibe-coding agencies are emerging Yes, actual agencies are forming around these tools. They're speeding up delivery, reducing costs, and unblocking indie builders and founders.
21. Conclusion 1. The European bias is real, Lovable vs. Bolt vs. Replit 2. Two types of users emerge 3. Cursor is a game-changer 4. Production isn’t the goal (yet) 5. Vibe-coding agencies are emerging
22. One more thing.
23. Introducing ****.ai beta
24. Introducing ****.ai beta Join the waitlist.
25. Introducing ****.ai beta Join the waitlist.
26. Thank you! pierre@strapi.io