The Startup Ideas Podcast

Episode Title: I Ranked Every Vibe Coding App (Cursor vs Claude Code vs Lovable)
Host: Greg Isenberg
Guests: Mickey (B), Unknown Third Voice (C)
Date: November 3, 2025

Episode Overview

In this engaging and opinionated episode, Greg Isenberg and guest Mickey present what they call the "definitive guide" and "ultimate tier list" for Vibe coding apps—a new wave of AI-powered coding platforms aimed at making app and software development radically easier. They compare and contrast top contenders like Cursor, Claude Code, Lovable, V0, Bolt, Codex, Windsurf, Roark, and Replit, delivering candid takes from both technical and non-technical perspectives. The conversation mixes technical analysis, industry gossip, and strategic advice for builders, culminating in actionable recommendations for listeners.

Key Discussion Points & Insights

1. Disclosures and Biases

• Neither Greg nor Mickey are investors in the companies spotlighted, though they clarify minor connections and personal biases.
• Mickey is employed at Convex, a backend platform discussed in the show.

2. How They Approach the Tier List (00:10–01:09)

• The list is based on personal experience and the hosts’ "non-professional professional opinion."
• Focus is on developer trust, accessibility for non-technical users, and technical robustness.

3. App-by-App Breakdown

Windsurf: The Fall from Grace

• [01:21 – 02:26]
• Quote:
  “Is there any trust anymore?...I probably would never use Windsurf again. And respectfully, in the kindest of fashion, I would put a D tier.” — Mickey (01:25)
• Reasons for Low Ranking:
  • Technical competence, but founder and team instability undermined trust.
  • Highly technical, not beginner-friendly.
• Community:
  • Built for developers, little adaptation for non-technical users.
  • Example of how team dynamics can impact a tool’s perceived reliability.

Cursor: Queen of the Hill for Devs

• [03:45 – 05:26]
• Quote:
  "If you are a vibe coder and non-technical person, you’re going to experience pain." — Greg (03:59)
  "I’ll give Cursor an A tier. I can agree with that. But to me, this is like S." — Mickey (05:26)
• Strengths:
  • Deep tutorial and documentation support.
  • Large, active community—similar to the rise of React in web dev.
  • Technical learning curve, but the premier choice for coders.
• Weaknesses:
  • Can be intimidating for non-technical people.

Lovable & Bolt: Easy but Restrictive

• [06:34 – 07:59 / 08:47 – 09:15]
• Quote:
  "The pain point with Lovable and Bolt was really the integrations...they’ve abstracted all this stuff away." — Mickey (06:43)
  "From a prototyping perspective...V0 is better than Lovable." — Mickey (07:53)
• Strengths:
  • Smooth onboarding for beginners via abstraction of integrations and backend.
• Weaknesses:
  • Limited backend choices (locked to Supabase).
  • Less flexibility compared to competitors.
• Tier Decision:
  • B (Greg prefers A/B between)
  • Bolt placed on par with Lovable.

V0: Built with Scale in Mind

• [07:57 – 09:47]
• Quote:
  "The integration with Vercel, it’s one of those things...it seems like Vercel and V0 are going to be very tight knit and that’s like a bet I’m willing to make." — Mickey (09:19)
• Strengths:
  • Versatile marketplace and backend options.
  • Tight integration with Vercel, a major industry player.
  • Robust templates and components.
  • Non-technical friendly.
• Tier Decision:
  • Between A and S for non-coders.

Mobile Vibe Coding Tools: Roark, Vibe Code app, Anything, Zero

• [10:21 – 14:18 / 13:34 – 14:50]
• Quote:
  "Vibe code app is...the lovable for mobile apps." — Greg (11:05)
• Trends & Insights:
  • Explosive growth in mobile-focused Vibe code apps; largely built using Expo and React Native.
  • Many are nascent but have huge upside, especially for TikTok-driven launches.
  • B tier for most current entrants (Roark, Anything, Zero).
  • Host tips: try free tiers, and expect rapid improvement.

Codex: The Most Improved

• [15:07 – 16:09]
• Quote:
  "My take on Codex is...it’s a bit. Yeah, it’s a little worse than cursor but it’s gained the most, most improved award." — Greg (16:09)
• Strengths:
  • Rapid improvement; close competition to Cursor and Claude.
  • Web version accessible to non-developers.
• Weaknesses:
  • Still not as refined as Cursor for technical users.

Claude Code: Once the King, Now ‘Nerfed’

• [16:53 – 18:28]
• Quote:
  "Recently has got nerfed. Yes. Like if this was 4 weeks ago...Claude code would be above S tier. But like now, I don’t know what’s happening." — Mickey (17:01)
• Strengths:
  • Formerly industry-leading. Runs on Claude Sonnet 4 or 5 models.
• Weaknesses:
  • Recent updates have dulled its edge.
  • Cursor seems to implement the same model but executes better agentically.
  • Expect a tug-of-war with Cursor for "best-in-class."
• Technical Deep Dive:
  • Output quality strongly influenced by platform agents and tools layered atop base models.

Replit: Powerful but Somewhat Opaque

• [18:28 – 20:10]
• Quote:
  "I don’t know anyone who’s used replit but they’re winning. I’m not even trying to be funny." — Mickey (18:37)
• Strengths:
  • Significant funding and user traction.
  • Two user modes: ‘vibe code’ and ‘agent’, catering to both fast prototyping and deeper builds.
  • Impressive outputs observed in some community demos.
• Weaknesses:
  • Mixed anecdotes, some find it less accessible.
  • Host hesitancy due to lack of direct experience.

Chef & Platform Ecosystem Notes

• [20:25 – 21:40]
• Chef (built by Convex) exists to showcase Convex’s backend power, not as a direct platform competitor.
• Many emerging Vibe code platforms are now building on Convex.
• Chef is open-sourced and not meant to impress end-users but to inspire developers.

4. Advice for Technical vs. Non-Technical Founders

• [21:40 – 24:33]
• Quote:
  "If you’re technical, then, man, it’s like Cursor or Cloud code...Now to non-technical people, I think there has to be...a mindset shift. A lot of people get frustrated if in five prompts they don’t build something...the audacity to think you’re going to build software...in 5, 6 prompts...it takes time, it takes planning, it takes testing. Software is art." — Mickey (21:50–23:50)
• For Technical People:
  • Cursor and Claude Code are the clear winners.
  • You’ll likely want full control and advanced features.
• For Non-Technical People:
  • Mindset shift required: real software always takes iteration, learning, and patience. There's no magic “5 prompt” solution to a finished product.
  • Emphasize documentation, community support, and using platform-specific planning tools (like Cursor’s Plan Mode).
  • Embrace the journey; mistakes and re-tries are part of the creative process.

5. Final Strategic Tip: Bet on Founders

• [24:33 – 25:13]
• Quote:
  “My recommendation to people is actually to follow all the founders of these businesses...Follow them and then see who you believe in...you are betting on someone, you’re betting on a company.” — Greg (24:33)
• Why:
  • Platforms are fluid, and teams’ vision/execution matter at least as much as tech stack.
  • Use founder communication, values, and track record as part of your selection criteria.

Notable Quotes & Timestamps

• “Is there any trust anymore?...I probably would never use Windsurf again. And respectfully, in the kindest of fashion, I would put a D tier.” — Mickey (01:25)
• “If you are a vibe coder and non-technical person, you’re going to experience pain.” — Greg (03:59)
• "I’ll give Cursor an A tier. I can agree with that. But to me, this is like S." — Mickey (05:26)
• "From a prototyping perspective...V0 is better than Lovable." — Mickey (07:53)
• "Vibe code app is...the lovable for mobile apps." — Greg (11:05)
• "My take on Codex is...it’s a bit. Yeah, it’s a little worse than cursor but it’s gained the most, most improved award." — Greg (16:09)
• "Recently has got nerfed. Yes. Like if this was 4 weeks ago...Claude code would be above S tier. But like now, I don’t know what’s happening." — Mickey (17:01)
• "I don’t know anyone who’s used replit but they’re winning. I’m not even trying to be funny." — Mickey (18:37)
• “If you’re technical, then, man, it’s like Cursor or Cloud code...Now to non-technical people, I think there has to be...a mindset shift. A lot of people get frustrated if in five prompts they don’t build something...Software is art.” — Mickey (21:50–23:50)
• “...you are betting on someone, you’re betting on a company, right?...people who bet on Windsurf got windsurfed.” — Greg (25:07–25:13)

Tier List Summary (Approximate)

| Tier | Platforms | |----------|---------------------------------------------------------------------------------------| | S / A | Cursor, V0 (for non-coders), (Claude Code formerly S, now A/B) | | A | Codex (most improved), Replit, Lovable* (Greg: A/B), Bolt | | B | Lovable* (Mickey: B), Bolt, Roark, Vibe code app (mobile), Anything, Zero (mobile) | | D | Windsurf |

(*Lovable's exact ranking is debated as A/B.)

Conclusion & Recommendations

• Best for Developers: Cursor, Claude Code (when stable), Codex (for those willing to ride its improvement wave).
• Best for Non-Technical Users: V0, Lovable, Bolt, Replit, Roark, and other mobile-focused tools—understanding that patience and persistence is essential.
• Mindset Matters: "Software is art"—expect an iterative journey, not instant magic.
• Follow the People: The founders and their visions shape these products; consider whose approach resonates with you.

Useful Segment Timestamps

• Windsurf discussion: [01:21–03:21]
• Cursor vs Windsurf: [03:45–05:26]
• Lovable/Bolt/V0: [06:34–09:47]
• Mobile app tools: [10:34–14:50]
• Codex: [15:07–16:20]
• Claude Code: [16:53–18:28]
• Replit: [18:28–20:10]
• Developer vs. Non-developer advice: [21:40–24:33]
• Bets on founders: [24:33–25:13]

For more free startup ideas, visit: gregisenberg.com/30startupideas