Lovable
First met in aicp-c3-basecamp-02 · Vibe-coding platforms
LovableFirst met in aicp-c3-basecamp-02 · Vibe-coding platforms
Describe an app in a browser chat and get a working, hosted, full-stack build — the course's default MVP factory.
You want the best-looking first draft fastest and you'll graduate the code out later — vs Bolt (weaker), Replit (more IDE), or Base44 (faster but less polished).
Won the bake-off on design — 'the best-looking output' across five platforms tested live in BC4
One-click GitHub sync, so the project can be cloned and taken over by a coding agent
Auth, secrets prompts and a code view are built in — it scaffolds a real app, not a mockup
Scaffolds a full app from a prompt: front end, routing, auth with email validation, database wiring
Pushes to a GitHub repo you own, so nothing is trapped in the platform
Prompts for the API keys a feature needs (OpenAI key for the AI project generator)
Shows the generated code, so you can learn from it or hand it to Claude Code
Doesn't produce a distinctive look by default — everything ships as the same purple gradient unless you name Tailwind/shadcn or feed it a reference screenshot
Doesn't stay cheap — 'a token eater'; free credits ran out inside one session
Doesn't wire cleanly to Supabase anymore — it steers you to Lovable Cloud instead
The first 20% of a build. Take the MVP to the point where the idea is proven, then clone the repo and continue in Claude Code or Cursor where iteration is cheaper.
Vibe-coding platforms
The KB is static HTML by design; no app to vibe-code yet.
The realistic candidate if the KB ever needs an interactive front end (search UI, review queue) beyond what build_site.py can emit — but the graduation-to-Claude-Code pattern matters more than the platform.
current
lovable-cloud: its bundled back end, pushed hard in place of Supabase
base44: faster to first build; Lovable won on design
claude-code: the documented graduation path — clone the Lovable repo and take over
supabase: the manual back end Lovable now discourages
| Session | Coverage | Moment | What happened |
|---|---|---|---|
| Basecamp 2: Vibe Coding | demonstrated | 1:08:32 | Built 'Task AI' end-to-end: scaffold, auth with email validation, secrets prompts, code view, GitHub connect, credit billing view; declined-then-adopted Lovable Cloud |
| Basecamp 4: Introduction to n8n — II | demonstrated | 2:17:35 | Bake-off winner on design; 'a token eater' — free credits nearly exhausted; repo pushed to GitHub and shared |
| Session 3: Building Elegant Landing Pages | demonstrated | 1:48:01 | Code-zip download demoed for migration to Codex; endorsed for beginners/prototyping with one-click hosting |
| Session 4: How to Build Micro Prototypes | demonstrated | 0:28:30 | Directory app one-shot from attached CSV, plan mode, AI-assistant mode via Lovable Cloud (model/db/secrets invisible), publish flow, code-zip download, GitHub integration; credit economics dissected |
| Session 16: GitHub 101 — Fundamentals & Essentials | demonstrated | 1:14:55 | Built the Sherlock escape-room app ('The Great Escape Society' / 'Baker Street Disappearance'), connected to GitHub via OAuth, auto-committed 12 times; also generated the README on the dev branch |
| Learning Topics: Sessions 1–5 — Foundations to AI Director | explained | The live CRM case study: planning mode, Lovable Cloud backend, per-user auth, publish failure debugged, security scan | |
| Workbooks 1–5: Guided Builds with Completion Proof | explained | Workbook 2's guided builder option (vs Bolt.new as the fast option) | |
| Session 14: Claude Code — Advanced Workflows | mentioned | 2:03:54 | The comparison baseline the shadcn skill beats on UI cleanliness |
| Session 15: Loop, Harness & Graph Engineering (posted as 'Securing Vibe-Coded Apps') | mentioned | 0:15:44 | 'Peanuts' next to Claude Code — the beginner rung, with Bolt |
| Session 17: Facebook Ads Competitor Research & Replication | mentioned | 0:50:47 | Named as the UI wrapper if the workflow were productized fully |
| Beginner Glossary (30 terms) + I'm Stuck Guides (23) | mentioned | The Lovable/Bolt-vs-Claude-Code guide: continue with the tool that fits; export a working version before major changes | |
| Practice Lab: Six Confidence-Building Projects | mentioned | Named among the builder options for projects 1–2 (with Bolt, Replit, Claude Artifacts) | |
| The Five-Level AI Generalist Roadmap (+ Operating System & Monetization) | mentioned | L5 non-technical track with Bolt, v0, Replit; also the micro-SaaS prompt-first stack | |
| Chapter 1: The Mindset — Figma Is a Sandbox, and Your Brain Is the Obstacle | mentioned | AI builders named and explicitly placed out of course scope | |
| AI Sprint: Claude Code with Antigravity — Day 2 (CLAUDE.md, the 50% Context Rule, and the Sketch-to-App Build) | mentioned | The named inspiration, treated respectfully: 'don't expect it to work exactly like Lovable' | |
| AI Sprint: The Hermes Agent — Day 2 (Loop Engineering: the Scheduling Loop, the Delegation Loop, and the Skill Auditor) | mentioned | Setup for the vibe-coding awareness story — the room outside doesn't know the words | |
| AI Masterclass: Exploring Antigravity 2 — Four Products, Four Slash Commands, and Two Live Builds | mentioned | His 'only good for mockups' claim — flagged as partisan; the corpus's Lovable projects say otherwise | |
| AI Sprint: The Open Source Playbook — Day 2 (Weights 101, OpenCode, and the Subsidy Math) | mentioned | Named as one of Anthropic's biggest B2B customers — the subsidy math's anchor | |
| AI Sprint: Windsurf & Devin — Day 1 (Vibe to Agentic: GitHub as Pizza, the Four Spec Concerns, and the IDE Lineage) | mentioned | Named in the vibe-coding conversation tier with Bolt and Emergent — descriptive, not dismissive |