FREE Cursor Built It in 8 Minutes... But Made a Mistake Copilot Didn't

Same spec, same prompt, same 22 automated checks that GitHub Copilot faced in Part 1. Cursor's free plan built the whole app in 8 minutes 12 seconds, wrote its own .gitignore, tested its own API, and scored 20/20. But reading the code, I found 3 red flags, including one thing Copilot got right and Cursor didn't. Everything runs on a 2016 dual-core laptop. $0 spent. I never touched the code. Prompt, test scripts, results and the exact code Cursor wrote: https://github.com/alphaadev/free-ai-... Chapters 0:00 Cursor vs Copilot: same prompt 0:24 Can free Cursor build a production app? 0:29 The series + Part 1 recap 0:56 The test machine ($0) 1:14 The app: NoteVault 1:38 The exact prompt 2:02 The rules 2:21 Scoring (locked before the build) 2:37 The build starts 3:06 Cursor's one-line plan 3:33 Timelapse: 23 files in 3 minutes 4:05 Cursor catches its own bug 4:32 npm install + docker compose up 5:02 Cursor tests itself 5:31 Done in 8 min 12 s 5:51 Minute-by-minute timeline 6:20 My tests: API contract + restart 6:50 All 22 checks 7:05 Browser test: sign up + notes 7:36 Browser test: the payment flow 8:10 UI checklist (no asterisk) 8:33 Under the hood: what it got right 9:09 3 red flags 10:02 Final score 10:28 Copilot vs Cursor scoreboard 10:52 Check these 3 things before you deploy 11:18 Next: Gemini CLI Series: Free AI Tools vs Production Part 1: GitHub Copilot | Part 2: Cursor | Next: Gemini CLI, Cline, Bolt #Cursor #AICoding #DevOps