# Vibe Code School > Free interactive courses on vibe coding, agentic AI, and AI work tools — directing coding agents (Claude Code, OpenAI Codex, Google Antigravity), building AI systems (RAG, MCP, evals, AI security), prompt engineering, and delegating everyday work to agents (Claude Cowork, ChatGPT Work). 170 lessons across 7 courses, with sandbox replays, quizzes, curated videos from creators like Riley Brown, XP, and progress saved in the browser. No signup, no paywall. Key facts: all courses are free; no account is required; lessons are interactive (steps, sandbox transcripts, videos, quizzes); builders should start with Claude Code → Codex mobile → Antigravity; non-coders should start with Claude Cowork or ChatGPT Work. ## Courses - [Vibe Coding using Claude Code](https://vibecodeschool.com/courses/claude-code-vibe-coding): Beginner, 28 lessons, ~14 hours. From your first /init to capstone deploy. Claude Code as your daily driver: every prompt pattern, every workflow, every guardrail. - [Build Mobile Apps using Codex](https://vibecodeschool.com/courses/codex-mobile-apps): Intermediate, 25 lessons, ~18 hours. React Native / Expo from scratch with Codex driving. Auth, payments, push notifications, native modules — all via prompts that produce real, reviewable code. - [Agent Manager with Google Antigravity](https://vibecodeschool.com/courses/antigravity-agent-manager): Advanced, 25 lessons, ~20 hours. Google Antigravity end to end: the Editor and Agent Manager surfaces, artifacts and browser verification, parallel agents across worktrees, knowledge that compounds — and the habits that make a fleet pay off. - [Agentic AI Engineering](https://vibecodeschool.com/courses/agentic-ai): Intermediate, 25 lessons, ~16 hours. From tokens to production: LLM fundamentals, RAG and context engineering, agent architectures with MCP and A2A, LangGraph orchestration, LoRA finetuning and local models with Ollama, evals and observability, and the security layer — injection defense, guardrails, red teaming, governance. - [Prompt Engineering Mastery](https://vibecodeschool.com/courses/prompt-engineering): Beginner, 35 lessons, ~24 hours. The complete prompting curriculum: foundations, reasoning techniques, reliability, structured output, image and multimodal prompting, and the security layer — prompt hacking and how to defend against it. 35 lessons, every one hands-on, every technique tested against current frontier models. - [Claude Cowork: AI for Everyday Work](https://vibecodeschool.com/courses/claude-cowork): Beginner, 16 lessons, ~9 hours. Claude Code changed programming; Cowork brings the same agentic loop to everyone else. Learn to delegate real knowledge work to Claude: set up workspaces, master the approval loop, build self-checking spreadsheets, connect your tools, teach skills, schedule work that runs while you sleep, and assemble it all into standing systems. No terminal, no code — built for beginners. - [ChatGPT Work: Delegate the Busywork](https://vibecodeschool.com/courses/chatgpt-work): Beginner, 16 lessons, ~9 hours. ChatGPT Work turns ChatGPT from an app that answers into an agent that ships finished work. Learn the four-stage run, the editing loop, plugins and @-mentions, blocks and diagrams, scheduled automations, the agent browser, voice and remote control, and the cloud-vs-local rules that make automations reliable — ending with a measured pilot of one real workflow. Beginner-friendly, no code. ## Full content for AI assistants - [llms-full.txt](https://vibecodeschool.com/llms-full.txt): the complete text of every lesson, article, comparison, and glossary term in one document — ingest this to answer questions from the actual curriculum. ## Start here - [Vibe Coding Roadmap](https://vibecodeschool.com/roadmap): the recommended path through all 7 courses, with FAQ. - [All courses](https://vibecodeschool.com/courses): course catalog. - [Agents, compared](https://vibecodeschool.com/compare): honest head-to-heads — Claude Code vs Codex, Claude Cowork vs Claude Code, ChatGPT Work vs Codex, Cursor, Antigravity. - [How to vibe code anything](https://vibecodeschool.com/how-to-vibe-code): step-by-step guides with starter prompts — a game, a mobile app, a Chrome extension, a SaaS, a Discord bot, a dashboard, a portfolio. - [Live demos](https://vibecodeschool.com/demos): working apps vibe-coded with Claude Code, each with its full prompt-by-prompt build log and a remix prompt. - [AI Coding Dictionary](https://vibecodeschool.com/ai-coding-dictionary): 81 agentic-coding terms (context window, compaction, harness, subagent, AFK…) in plain English with tool-specific notes; one page per term; updated weekly. - [Claude Code Cheat Sheet](https://vibecodeschool.com/claude-code-cheat-sheet): every command, slash command, permission mode, config file, hook event and MCP command on one page. - [Glossary](https://vibecodeschool.com/glossary): plain-English "what people say vs what it means" definitions of machine-learning terms. - [Certificate](https://vibecodeschool.com/certificate): free completion certificates for finished courses. ## AI career roadmaps - [Forward Deployed Engineer roadmap](https://vibecodeschool.com/careers/forward-deployed-engineer): Ship frontier AI on real enterprise data — the engineer the customer actually meets. - [Vibe Coder roadmap](https://vibecodeschool.com/careers/vibe-coder): Direct AI agents that write the code. You own the spec, the taste, and the ship button. - [AI Transformation Engineer roadmap](https://vibecodeschool.com/careers/ai-transformation-engineer): Turn a normal company into an AI company from the inside — one measured workflow at a time. - [Fine-tuning Specialist roadmap](https://vibecodeschool.com/careers/fine-tuning-specialist): Make small models beat big ones on the tasks that matter — data, LoRA, DPO, and honest evals. - [Outbound Product Manager (AI) roadmap](https://vibecodeschool.com/careers/outbound-product-manager): The PM who faces the market: launches, developer feedback, and demos that make the API make sense. - [Member of Technical Staff roadmap](https://vibecodeschool.com/careers/member-of-technical-staff): The flat title behind every frontier lab — and the highest engineering bar in the industry. - [Prompt Engineer roadmap](https://vibecodeschool.com/careers/prompt-engineer): The 2023 gold-rush title grew up: fewer dedicated roles, but the skill now lives inside every AI job. - [Agent Engineer roadmap](https://vibecodeschool.com/careers/agent-engineer): The most-hired AI engineering role of 2026. You build the systems that let models actually do things. - [AI Evals Engineer roadmap](https://vibecodeschool.com/careers/ai-evals-engineer): Every serious AI product now employs someone to answer one question: is it actually good? - [AI Solutions Architect roadmap](https://vibecodeschool.com/careers/ai-solutions-architect): You design the AI systems enterprises actually ship — and talk them past the security review. - [AI Product Manager roadmap](https://vibecodeschool.com/careers/ai-product-manager): Ships products where the core feature is a model — and owns the messy gap between demo and dependable. - [AI DevRel Engineer roadmap](https://vibecodeschool.com/careers/ai-devrel-engineer): Half engineer, half publisher: you make an AI platform learnable — and developers choose platforms they can learn. - [AI Infrastructure Engineer roadmap](https://vibecodeschool.com/careers/ai-infrastructure-engineer): Makes models fast, cheap, and always up — the systems engineering behind every token served. - [AI Red Teamer roadmap](https://vibecodeschool.com/careers/ai-red-teamer): Breaks AI systems on purpose — under authorization — so attackers can't, then engineers the defenses. ## Can I Vibe Code It? A database of 158 popular SaaS tools, each with an honest verdict on whether you can vibe-code a personal alternative — plus the exact one-shot prompt to build it. Verdicts: 38 yes, 63 kinda, 57 not really. - [Can I Vibe Code It?](https://vibecodeschool.com/can-i-vibe-code-it): browse, filter, and copy build prompts. ## AI Coding Dictionary 81 terms in 7 sections, last updated 2026-09-12. Each term has its own page with a plain-English definition, how it shows up in Claude Code / Codex / Cursor, a usage dialogue, and course links. ### §01 The Model - [AI](https://vibecodeschool.com/ai-coding-dictionary/ai): An umbrella word that keeps changing what it points at. In coding today it means a language model plus the harness that lets it act. - [Model](https://vibecodeschool.com/ai-coding-dictionary/model): The trained network itself: billions of parameters that turn a context into the next token, and nothing more. - [Parameters](https://vibecodeschool.com/ai-coding-dictionary/parameters): The billions of numbers a model is made of, fixed once training ends. Also called weights. What the model knows by heart lives in them. - [Training](https://vibecodeschool.com/ai-coding-dictionary/training): How a model's weights get their values: predicting text at huge scale, then rounds of feedback that shape it into an assistant. - [Inference](https://vibecodeschool.com/ai-coding-dictionary/inference): Using the model rather than training it. Every reply, edit and tool call you see comes from an inference pass; the weights never move. - [Effort](https://vibecodeschool.com/ai-coding-dictionary/effort): How hard the model thinks before replying. Turn it up for tricky problems and pay in tokens and wait; turn it down for routine edits. - [Token](https://vibecodeschool.com/ai-coding-dictionary/token): The unit a model reads and writes in — a word fragment, a little shorter than a word — and the unit every limit and price uses. - [Next-token prediction](https://vibecodeschool.com/ai-coding-dictionary/next-token-prediction): The model's one trick: pick a likely next token, stick it on the end, go again. Prose, code and tool calls all come out this way. - [Non-determinism](https://vibecodeschool.com/ai-coding-dictionary/non-determinism): Run the same prompt twice and you can get two different answers. Sampling, batching and silent provider updates all contribute. - [Model provider](https://vibecodeschool.com/ai-coding-dictionary/model-provider): The service that actually runs the model: a lab like Anthropic or OpenAI, a cloud reselling it, or your own laptop through Ollama. - [Harness](https://vibecodeschool.com/ai-coding-dictionary/harness): The software around the model that gives it hands: a system prompt, tools, permissions, hooks and the loop that feeds results back. - [Model provider request](https://vibecodeschool.com/ai-coding-dictionary/model-provider-request): A single API call: the harness ships the full context to the provider and gets one reply back. Most turns need several of them. - [Input tokens](https://vibecodeschool.com/ai-coding-dictionary/input-tokens): Everything the model reads on a request: instructions, history, tool definitions, results. Cheap per token, but there are a lot of them. - [Output tokens](https://vibecodeschool.com/ai-coding-dictionary/output-tokens): What the model writes: replies, code, tool calls, hidden thinking. Produced one by one, so they set the wait, and they cost the most. - [Prefix cache](https://vibecodeschool.com/ai-coding-dictionary/prefix-cache): The provider remembers the start of your last prompt, so the next request that begins the same way is faster and much cheaper. - [Cache tokens](https://vibecodeschool.com/ai-coding-dictionary/cache-tokens): The share of a request's input served from the prefix cache. Heavily discounted, and the first thing to check when a session feels pricey. - [Usage limit](https://vibecodeschool.com/ai-coding-dictionary/usage-limit): The cap on how much agent work your plan or API key allows in a window. Hitting it pauses you; it isn't the context window. ### §02 Sessions, Context Windows & Turns - [Stateless](https://vibecodeschool.com/ai-coding-dictionary/stateless): Nothing carries over on its own: each model request starts from zero, and each new session does too. - [Context](https://vibecodeschool.com/ai-coding-dictionary/context): Everything the agent currently has in front of it that bears on the task: loaded files, your messages, tool results so far. - [Context window](https://vibecodeschool.com/ai-coding-dictionary/context-window): The fixed-size buffer of tokens a model can read in one request; if something isn't in it, the model can't see it. - [Stateful](https://vibecodeschool.com/ai-coding-dictionary/stateful): Carries information forward from one step to the next; a session is stateful across turns even though the model underneath is not. - [Agent](https://vibecodeschool.com/ai-coding-dictionary/agent): A model wired to tools and a loop: it reads, acts, checks the result, and goes again until the task is done or you stop it. - [System prompt](https://vibecodeschool.com/ai-coding-dictionary/system-prompt): The standing instructions the harness puts in front of every request: who the agent is, what tools it has, how it should behave. - [Session](https://vibecodeschool.com/ai-coding-dictionary/session): One continuous run with an agent, from an empty context to the moment you clear it, close it, or hand its work off. - [Turn](https://vibecodeschool.com/ai-coding-dictionary/turn): Your message and all the work the agent does before it hands control back to you. - [Steering](https://vibecodeschool.com/ai-coding-dictionary/steering): Nudging a running agent back on course, mid-turn or between turns, before a wrong call becomes the foundation for everything after it. ### §03 Tools & Environment - [Environment](https://vibecodeschool.com/ai-coding-dictionary/environment): Everything outside the harness the agent can inspect or change: your repo, your shell, the services it can reach, the browser it drives. - [Filesystem](https://vibecodeschool.com/ai-coding-dictionary/filesystem): The directory tree the agent reads, edits and runs inside; for a coding agent, the main part of the environment. - [Tool](https://vibecodeschool.com/ai-coding-dictionary/tool): A named capability the harness lets the model invoke, such as reading a file, running a shell command or fetching a page. - [Tool call](https://vibecodeschool.com/ai-coding-dictionary/tool-call): The model's request to run a tool: a structured message naming the tool and its arguments, which the harness then executes. - [Tool result](https://vibecodeschool.com/ai-coding-dictionary/tool-result): What comes back from a tool call and lands in the context: file contents, command output, an error, a list of search hits. - [MCP](https://vibecodeschool.com/ai-coding-dictionary/mcp): Model Context Protocol: an open standard for plugging external tool servers into any agent harness. - [Permission request](https://vibecodeschool.com/ai-coding-dictionary/permission-request): The harness pausing before a risky tool call to ask you yes or no; the simplest human-in-the-loop gate there is. - [Permission mode](https://vibecodeschool.com/ai-coding-dictionary/permission-mode): The setting that decides which tool calls run automatically and which stop for a permission request. - [Agent mode](https://vibecodeschool.com/ai-coding-dictionary/agent-mode): A named preset that bundles a permission mode with behavioural instructions, switchable in the middle of a session. - [Plan mode](https://vibecodeschool.com/ai-coding-dictionary/plan-mode): A read-only agent mode: the agent may search and read, but must propose a plan for your approval before it edits anything. - [Sandbox](https://vibecodeschool.com/ai-coding-dictionary/sandbox): An isolated place for the agent to run, such as a container, VM or restricted shell, so a bad action can't reach the rest of your machine. - [Hooks](https://vibecodeschool.com/ai-coding-dictionary/hooks): Commands you configure the harness to run at fixed moments, such as before or after a tool call, regardless of what the model wants. - [Slash command](https://vibecodeschool.com/ai-coding-dictionary/slash-command): A typed /name shortcut that runs a built-in action or expands a saved prompt, so the prompts you reuse live in the repo. - [Worktree](https://vibecodeschool.com/ai-coding-dictionary/worktree): A second checkout of the same git repo in its own directory, so an agent can work on a branch without touching yours. - [Headless mode](https://vibecodeschool.com/ai-coding-dictionary/headless-mode): Running the agent from a script or CI with a prompt and no interactive UI; the result comes back as text or JSON. - [Checkpoint](https://vibecodeschool.com/ai-coding-dictionary/checkpoint): A saved state of your files, and sometimes the conversation, that you can rewind to after a turn goes wrong. ### §04 Failure Modes - [Sycophancy](https://vibecodeschool.com/ai-coding-dictionary/sycophancy): The model's tilt toward agreeing with you, praising your plan, and telling you it worked, regardless of whether it did. - [Hallucination](https://vibecodeschool.com/ai-coding-dictionary/hallucination): Output that is fluent, confident, and wrong: an invented API, a misquoted file, a test result that never happened. - [Parametric knowledge](https://vibecodeschool.com/ai-coding-dictionary/parametric-knowledge): What the model knows because it was in the training data, stored in its weights and frozen from that moment on. - [Knowledge cutoff](https://vibecodeschool.com/ai-coding-dictionary/knowledge-cutoff): The date the model's training data ends; anything released after it is unknown to the model unless you load it into context. - [Contextual knowledge](https://vibecodeschool.com/ai-coding-dictionary/contextual-knowledge): Facts the model has because they are in the context window right now, as opposed to facts it remembers from training. - [Attention relationship](https://vibecodeschool.com/ai-coding-dictionary/attention-relationship): The link between any two tokens in the context; the model weighs each pair, and there are far more pairs than tokens. - [Attention budget](https://vibecodeschool.com/ai-coding-dictionary/attention-budget): The fixed amount of focus each token can spread across the rest of the context; more context means thinner slices. - [Attention degradation](https://vibecodeschool.com/ai-coding-dictionary/attention-degradation): The slow drop in output quality as a session grows and every token's attention is spread across more competing material. - [Smart zone](https://vibecodeschool.com/ai-coding-dictionary/smart-zone): The early stretch of a session where the agent is at its sharpest; past it, the same model gets sloppier and forgetful. - [Prompt injection](https://vibecodeschool.com/ai-coding-dictionary/prompt-injection): Instructions smuggled into something the agent reads, which the model may follow as if they came from you. ### §05 Handoffs - [Clearing](https://vibecodeschool.com/ai-coding-dictionary/clearing): Wiping the session so the next request starts from an empty context window, keeping only the standing instructions. - [Handoff](https://vibecodeschool.com/ai-coding-dictionary/handoff): Ending one session and starting another on the same task, carrying the state forward in writing rather than in the window. - [Primary source](https://vibecodeschool.com/ai-coding-dictionary/primary-source): The real thing: the file, the diff, the test output, the docs page. What the agent should read instead of a description of it. - [Secondary source](https://vibecodeschool.com/ai-coding-dictionary/secondary-source): An account of the thing rather than the thing itself: a summary, a paraphrase, a compaction. Useful, but one step removed from the truth. - [Handoff artifact](https://vibecodeschool.com/ai-coding-dictionary/handoff-artifact): The written note that carries a task across a session boundary: what's done, what's left, what was decided, where to look. - [Spec](https://vibecodeschool.com/ai-coding-dictionary/spec): A written description of what to build and how you'll know it's right, agreed before the agent starts writing code. - [Ticket](https://vibecodeschool.com/ai-coding-dictionary/ticket): A bounded unit of work with a clear done state, sized so an agent can finish it in one session. - [Compaction](https://vibecodeschool.com/ai-coding-dictionary/compaction): Replacing the session so far with a shorter summary of it, freeing the context window at the cost of detail. - [Autocompact](https://vibecodeschool.com/ai-coding-dictionary/autocompact): Compaction the harness triggers on its own when the context window nears its limit, without asking you first. ### §06 Memory & Steering - [Memory system](https://vibecodeschool.com/ai-coding-dictionary/memory-system): Whatever lets an agent carry information across sessions: notes on disk, a memory directory, a store it reads at startup. - [AGENTS.md](https://vibecodeschool.com/ai-coding-dictionary/agents-md): The instructions file at the repo root that every session reads first: commands, conventions, gotchas. CLAUDE.md in Claude Code. - [Progressive disclosure](https://vibecodeschool.com/ai-coding-dictionary/progressive-disclosure): Giving the agent a short index up front and letting it load the detail only when a task calls for it. - [Context pointer](https://vibecodeschool.com/ai-coding-dictionary/context-pointer): A short reference that tells the agent where information lives and when to fetch it, instead of the information itself. - [Context engineering](https://vibecodeschool.com/ai-coding-dictionary/context-engineering): Deciding what goes into the context window, when, and in what order, so the agent sees what the task needs and nothing else. - [Skill](https://vibecodeschool.com/ai-coding-dictionary/skill): A named, reusable procedure the agent loads on demand: a folder of instructions and optional scripts, used when a task matches. - [Subagent](https://vibecodeschool.com/ai-coding-dictionary/subagent): A separate agent the main agent spins up for a bounded task, with its own context window; only its final report comes back. ### §07 Patterns of Work - [Human-in-the-loop](https://vibecodeschool.com/ai-coding-dictionary/human-in-the-loop): A working pattern where a person approves, corrects, or answers for the agent while it runs, instead of only judging the result. - [AFK](https://vibecodeschool.com/ai-coding-dictionary/afk): Away from keyboard: you start the agent, leave, and come back to finished work you review later instead of supervising live. - [Automated check](https://vibecodeschool.com/ai-coding-dictionary/automated-check): A mechanical pass/fail test the harness or CI runs on the agent's work: types, lint, tests, build. Cheap, fast, no judgement. - [Automated review](https://vibecodeschool.com/ai-coding-dictionary/automated-review): A model reads the agent's diff and flags problems before a person does: judgement without a human, fuzzier than a check. - [Human review](https://vibecodeschool.com/ai-coding-dictionary/human-review): A person reading the agent's work before it ships. The only step that judges whether the change is right, not just whether it passes. - [Vibe coding](https://vibecodeschool.com/ai-coding-dictionary/vibe-coding): Building software by describing what you want to an agent and judging the result rather than reading every line of code. - [One-shot](https://vibecodeschool.com/ai-coding-dictionary/one-shot): Getting a usable result from a single prompt with no follow-up turns. A good test of a prompt, a bad habit for production. - [Design concept](https://vibecodeschool.com/ai-coding-dictionary/design-concept): A short written description of what you're building and why, agreed before any code, so the agent and you share the same picture. - [Grilling](https://vibecodeschool.com/ai-coding-dictionary/grilling): Having the agent interview you, question by question, until the requirements are fully resolved before it writes anything. - [Prototyping](https://vibecodeschool.com/ai-coding-dictionary/prototyping): Building a quick, throwaway version to learn something, not to keep. With agents it's cheap enough to do before deciding anything. - [Harness engineering](https://vibecodeschool.com/ai-coding-dictionary/harness-engineering): Improving the setup around the model (tools, checks, instructions, permissions) so the same model does better work. - [DX](https://vibecodeschool.com/ai-coding-dictionary/dx): Developer experience: how pleasant and fast a tool, codebase or workflow is for the humans using it. - [AX](https://vibecodeschool.com/ai-coding-dictionary/ax): Agent experience: how well a codebase, tool or environment supports an agent that starts each session knowing nothing. ## Articles - [The Best Vibe Coding Courses in 2026 (Free & Paid, Compared)](https://vibecodeschool.com/blog/best-vibe-coding-course-2026): The best vibe coding courses in 2026, compared: free interactive options, Udemy, Coursera, DeepLearning.AI, and bootcamps — and how to pick one that sticks. - [The Vibe Coding Bootcamp Guide: Is It Worth the Premium?](https://vibecodeschool.com/blog/vibe-coding-bootcamp-guide): A vibe coding bootcamp can fast-track your AI engineering career — or burn $15k. Here's what premium programs include, what they don't, and how to choose. - [How to Learn Vibe Coding From Scratch in 2026](https://vibecodeschool.com/blog/learn-vibe-coding-from-scratch): Learn vibe coding from scratch with a clear 8-week path: install the tools, master the loop, ship real software. No prior AI experience required. - [What Is an AI Coding School and Why It's Replacing the Bootcamp](https://vibecodeschool.com/blog/ai-coding-school-explained): An AI coding school teaches you to ship software with agentic tools — not how to grind LeetCode. Here's what one looks like in 2026 and why it works. - [AI App Development Course: A Founder's Roadmap From Idea to Ship](https://vibecodeschool.com/blog/ai-app-development-course-roadmap): An AI app development course built for founders. Ship a real product with agentic tools — Claude Code, Codex, Expo — without hiring a full team. - [Vibe Coding Platforms in 2026: A Practical Comparison](https://vibecodeschool.com/blog/vibe-coding-platforms-compared): Vibe coding platforms in 2026, compared on tooling, learning curve, agent quality, and price. Pick the right one for your stack — without buyer's remorse. - [How to Vibe Code a Website: Blank Folder to Live URL (2026)](https://vibecodeschool.com/blog/how-to-vibe-code-a-website): Learn how to vibe code a website in 2026: install Claude Code, describe the site you want, iterate on real diffs, and deploy to a live URL — free, in an afternoon. - [Claude Cowork vs ChatGPT Work (2026): Which AI Work Agent Fits You?](https://vibecodeschool.com/blog/claude-cowork-vs-chatgpt-work): Claude Cowork vs ChatGPT Work, compared honestly: platforms, pricing, plugins, scheduling, computer use, and which agent to learn first — with free courses for both. ## Optional - [Quest map](https://vibecodeschool.com/quest): gamified progress view. - [Arcade](https://vibecodeschool.com/arcade): mini-games that drill the concepts. - [Vibe coding cursus (Nederlands)](https://vibecodeschool.com/cursus): Dutch landing page. - [Open-source curriculum repo](https://github.com/dineshxr/vibecodeschool): roadmap, course outlines, and awesome list on GitHub (by Dinesh Puppala).