An AI-first code editor — a familiar VS Code surface with an agent living inside it.
Pick this whenYou want to stay in an editor and see every diff as it lands, rather than conversing with a terminal agent.
What makes it different
Ships its own fast model, so quick edits don't cost premium tokens
Free tier is usable as a plain editor — no subscription needed to try it
Editor-native: the agent works where you already read code
What it does
Edits code with AI assistance inside a full IDE
Runs agentic multi-file changes like its terminal peers
Works as an ordinary editor when you don't want the AI
What it does not do
Common wrong expectationsDoesn't justify a parallel subscription — the course's advice is to pick one local coding tool at a time
Where it sits in a stack
The same seat as Claude Code, for people who think in an editor rather than a terminal.
ShelfCoding agents & editors
Our status · not-usingClaude stack covers this seat.
For your projectsOnly relevant if the pipeline ever needs hands-on editing away from an agent session.
| Session | Coverage | Moment | What happened |
|---|
| Office Hours V: Paul's Compliance Scanner Meets Firecrawl, Painkillers vs Vitamins, What a Real Second Brain Costs, and Why Selling Got Harder | demonstrated | | Agent window vs editor window |
| Session 9: Introduction to Cursor & its principles | demonstrated | 0:33:40 | The session's platform: Composer 2.5 default model (built on Kimi K2.5), settings → rules/skills/MCP, fresh-agent workflow, @-file references — 'super fast' at build time; SpaceX/X acquisition news relayed |
| Session 10: Building on Cursor & connect with AI workflows | demonstrated | 0:19:21 | The build platform end-to-end: sub-agents, superpowers interviews, ~1M tokens across the session; Composer 2.5 implementing, Opus 4.8 suggested for planning mode |
| Session 16: GitHub 101 — Fundamentals & Essentials | demonstrated | 2:02:19 | Its agent ran the local half: terminal folder creation, git clone (102 files), three local branches, pushes, browser-based GitHub auth |
| Day 8: AI Workflows — Chaining Tools, the IDEA(L) Analysis Framework, Claude Skills vs MCP vs Projects, and the Reel-Cloning Pipeline | demonstrated | | WhatsApp MCP server installed by prompt |
| Day 1: What Makes an LLM an Agent - Planning, Tools, Memory, Actions; Five Design Patterns; and Cursor, Antigravity and Claude Code Live | demonstrated | | Ask / Agent / Plan modes |
| AI Sprint: Vibe Coding to Production — Day 1 (Deployment, DNS, and the Five Locks) | demonstrated | | His coding agent for the build + the audit prompt |
| AI Sprint: AI Evals & Reliability — Day 2 (Testing Systems That Never Answer Twice) | demonstrated | | Live-coding the smallest possible eval in Python |
| Building a Web App with Replit: Vibe Coding Explained, the Tool Landscape, and a Live Build of a Business Site with an AI Sales Chatbot | explained | | |
| AI Sprint: Meet GrokBot — Stop Prompting, Start Delegating (One Cloud Computer, a Chief of Staff, Four Specialist Bots, and the Reversible/Irreversible Rule) | explained | | Ultra plan bundles GrokBot; coding bots run Cursor Cloud Agents (as heard) |
| Basecamp 2: Vibe Coding | mentioned | 1:33:06 | Local alternative alongside Claude Code; has its own fast model; free tier usable as plain editor |
| Community Session II: The Meta-Harness — Praful Thakkar on AgentGrid, Orchestrator-plus-Workers on One Canvas, Hand-Off Between Models, and OAuth Instead of API Keys | mentioned | | Used before, not with AgentGrid |
| Office Hours VI: Paul's SEO/AEO/GEO Question - Moving a Lovable Site into GitHub and Claude Code, the GEO Skill, llms.txt, and CLIs as the Glue | mentioned | | Host's measured opinion |
| Office Hours VIII (platform label 'XIII'): App-Store Economics, Vetting a Co-Founder Without Your Code, Android Accessibility Agents, and OpenClaw vs Hermes | mentioned | | Per-sub-agent model assignment |
| Session 11: Overview of Claude Code & Key Concepts (Installs, Modes, Context Files, Skills, Hooks & MCPs) | mentioned | 0:26:49 | Asked about early; comparison deferred |
| Session 12: Claude Code — Design & Building Agents (App-First: a Game, a GTM Board, and a Live Deploy) | mentioned | 2:38:38 | Q&A comparison: both agentic coders; Cursor 'highly technical', Claude Code easier |
| Session 15: Loop, Harness & Graph Engineering (posted as 'Securing Vibe-Coded Apps') | mentioned | 0:15:44 | Placed 'maybe one above beginner' on the ladder |
| AI for Designers: Product Research - Personas, Jobs to Be Done, Competitors and a Locked Value Proposition in Claude and ChatGPT | mentioned | | |
| The Five-Level AI Generalist Roadmap (+ Operating System & Monetization) | mentioned | | L5 developer track; Composer 2.5 spotlight as the own-the-inference-stack example (dated benchmarks) |
| Learning Topics: Sessions 1–5 — Foundations to AI Director | mentioned | | Tier-two AI IDE on the builder ladder, with Windsurf |
| Day 2 (morning): Agentic Workflows in Code - CrewAI from One Agent to a DevOps Crew to a Parallel Investment Adviser with Custom Tools | mentioned | | Trainer's IDE |
| Vibe Coding for Non-Coders: Pain Point to MoSCoW to Stack, 'Use AI to Build AI' with Grok, and an Audio Note-Taker Built Live on Lovable + Supabase + OpenAI | mentioned | | Where he goes when Lovable hits its ceiling |
| AI Sprint: AGI Readiness - Day 1 (Three Claims and Three Dials, the Agent Harness and Autonomy Ladder, the Failure Mode of Abundance, and the Six-Layer Readiness Stack) | mentioned | | |
| AI Sprint: Windsurf & Devin — Day 1 (Vibe to Agentic: GitHub as Pizza, the Four Spec Concerns, and the IDE Lineage) | mentioned | | Fellow VS Code fork; Composer named as Cascade's counterpart |
| AI Sprint: Windsurf & Devin — Day 2 (Building Superplexity: Plan with Opus, Execute with Cascade, Review with Devin) | mentioned | | Cursor+Claude Code named as the equivalent IDE+agent pairing |
Cursor