Scenario Practice: Choose Your Route—Website, Mini Program, or Browser Extension
💡 Goal: In 5 minutes, choose your first practice route—website, Mini Program, or browser extension—and be able to state that route's barriers, approximate time cost, and the entry point to its first lesson.
Now that you can use the tools and have set the disciplines, the first question that truly trips people up in vibe coding practice is not "how to write," but "what to build." "Make me something" is the request AI handles worst; the more specific the goal, the more likely AI gets it right in one shot—this is the principle covered in 3.2 prompt-writing techniques. So this stop first tackles project topic selection: build a website, a WeChat Mini Program, or a browser extension with AI—pick one of the three routes, then get to work.
The three routes are divided by "where the finished product is used," with different barriers and launch requirements:
| Route | What the result is | Where users use it | What it takes to go live |
|---|---|---|---|
| A Website | Personal homepage / portfolio / landing page | Any browser, just open the link | Lightest: publish by uploading, get an accessible URL |
| B WeChat Mini Program | A small tool that spreads within WeChat | Acquired and shared inside WeChat | Medium: trial version needs no account or review; official release requires registration + filing + WeChat review |
| C Browser extension | A tool that enhances the browser | Install and use in Chrome | Personal use requires no review; listing in the store requires a registration fee + review |
Still hesitating over "can I build something without knowing how to code"? First go back to 1.3 the honest answer for beginners for real cases, then come back to pick a route—what turns "wanting to learn" into "finished" has always been a concrete goal, not the tool.
🎯 Stage Goals
- Be able to choose one of the three routes and state your reasons: who it's for and where it will be used
- Know the full course list, total time cost, and what must be cleared before launch (free publishing / filing and review / store review) for your chosen route, so you don't get stuck halfway through on eligibility requirements
- Start the first lesson that same day, no longer stuck at "want to learn but don't know what to build"
⏳ Estimated Time
- This page takes 5 minutes; pick one of the three routes and complete it:
- Route A (Website): 5 lessons totaling about 62 minutes, with a variance of 43–81 minutes
- Route B (Mini Program): 5 lessons totaling about 57 minutes, with a variance of 40–74 minutes
- Route C (Extension): 7 lessons totaling about 78 minutes, with a variance of 55–101 minutes
- You don't need to do all three: fully finishing one beats doing half of each
🚀 Before You Start
- Completed the previous stopEfficient Workflow—the "plan before you build" and "acceptance check before launch" disciplines from the five disciplines are used throughout practice
- At least one tool can run a full conversation round (if not installed, go back to 1.5 Installing Claude Code; if you really don't want to touch the command line, see 1.5b Alternative Installation Methods)
- Have a real small goal as practice material (a portfolio, a signup page, an operation that saves yourself a step)—a real goal lasts longer than a "practice project"
- A cup of water ☕
📚 Choose Your Route
The three routes aren't ranked; it's only about whether they fit your goal. The fastest way to choose is by looking at "where the users are": any browser → A; inside WeChat → B; mainly just you → C. Each card below lists the barriers and courses clearly—once you've chosen, jump straight into the first lesson.
Route A: Build a Website—Share a Link Anyone Can View
Who it suits: People who want a front door for themselves or their work—personal homepage, portfolio, event landing page, content site.
Why the barrier is lowest:
- All three hosting platforms have free tiers: GitHub Pages (publish a static site straight from a repository), Cloudflare Pages (drag-and-drop upload even without knowing Git), Vercel (Hobby plan is free)
- No server to buy: a static site is just HTML/CSS/JS files, and the platform builds and distributes it for you
- The shortest path shows results the same day: with Cloudflare Pages drag-and-drop upload, drag a folder or zip into the browser and you get a
项目名.pages.devURL
Things to note:
- GitHub Pages' free tier requires the repository to be public, and the site content is visible on the public web—setting the repository to private doesn't change this
- Vercel's Hobby plan is for personal, non-commercial use only; don't put revenue-generating projects on the free tier
- Cloudflare's official documentation now recommends new projects start with Workers; A3 still uses Pages drag-and-drop upload as the teaching path because it's the most direct for beginners
Course schedule (5 lessons, about 62 minutes):
| Lesson | Content | Duration |
|---|---|---|
| A1 Build a Website with AI | From one sentence to a launch-ready page | 15 minutes |
| A2 Launch on GitHub Pages | Free static website hosting | 15 minutes |
| A3 Launch on Cloudflare Pages | Drag-and-drop publishing even without Git | 12 minutes |
| A4 Launch on Vercel | One-command deployment from an AI coding agent | 12 minutes |
| A5 Deployment platform comparison | Free quotas, limits, and how to choose | 8 minutes |
Route B: Build a WeChat Mini Program—Make it spread inside WeChat
Who it's for: People whose users are on WeChat—build a small tool that clubs, shops, courses, or events can use with a simple QR scan.
Why it's worth doing:
- The official positioning is "a brand-new way to connect users with services," easily accessed and spread within WeChat
- Individuals can register: Any WeChat user aged 18 or older with real-name identity information in mainland China can do it—no company required
- There's a shortcut for beginners: WorkBuddy, as a WeChat third-party service provider, assists with code upload and trial version generation—the trial version requires no account registration in advance and no WeChat review, valid for 14 days with at most 15 testers
Things to note:
- Official public release takes a few extra steps: register a Mini Program account, complete the filing (ICP), and pass WeChat review
- Individual entities don't have WeChat Pay; to collect money you must register a business entity (300 CNY verification fee, or verification via corporate bank account transfer)
- A trial version can't be directly converted to the official version; to switch release methods, you must "unbind" in the sharing panel and start over
Course schedule (5 lessons, about 57 minutes):
| Lesson | Content | Duration |
|---|---|---|
| B1 What is a WeChat Mini Program | Opportunities beginners can seize | 8 minutes |
| B2 Registering a Mini Program account | Materials, process, and individual vs. business differences | 12 minutes |
| B3 Building a Mini Program with WorkBuddy | From idea to trial version in 14 days | 15 minutes |
| B4 Submitting for review and publishing | A guide to avoiding pitfalls in WeChat review rules | 12 minutes |
| B5 What to do if review is rejected | Common rejection reasons and fixes | 10 minutes |
Route C: Build a browser extension—modify your own browser
Who it's for: People who spend all day in Chrome and want to solve their own pain points while browsing the web—and can publish to users worldwide as a bonus.
Why choose it:
- The tech stack is just the classic web trio: in the official words, you can build extensions with the same technologies used to build web apps—HTML, CSS, JavaScript—if you can build a web page, you can build an extension
- Build and use instantly: turn on Developer mode on the Chrome extensions page, use Load unpacked to load a local folder and it runs—personal use requires no review at all
- You can do more than you imagine: toolbar icon popups, sidebars, intercepting or modifying network requests, injecting scripts into specific pages, listening to browser events in the background
Things to note:
- Only write for Manifest V3: MV2 extensions have been disabled across all Chrome channels since 2025-7; following old tutorials verbatim will be wasted effort
- Publishing to the Chrome Web Store requires a one-time registration fee (the amount is subject to the actual payment page in the developer dashboard); mainland China is on the list of supported registration regions
- Store review looks at three things: a single purpose, clear privacy disclosure, and all code bundled in the package—no remotely loading JS at runtime
Course schedule (7 lessons, about 78 minutes):
| Lesson | Content | Duration |
|---|---|---|
| C1 What is a browser extension | If you can build a web page, you can build an extension | 8 minutes |
| C2 Your first Chrome extension | Get Hello World running in 15 minutes | 15 minutes |
| C3 manifest.json explained | Quick reference for MV3 fields | 10 minutes |
| C4 Content Scripts and Service Workers | The extension's brain and hands | 12 minutes |
| C5 When your extension won't come up | Where to find logs and how to read errors | 10 minutes |
| C6 Publishing to the Chrome Web Store | The full process: registration, packaging, and submitting for review | 15 minutes |
| C7 Store review and release strategy | Review done in 3 days, staged rollout and rollback | 8 minutes |
💡 Still can't decide? Two questions determine your route
- Where are your users? In WeChat → B; in any browser → A; mainly just you using it every day → C
- How fast do you want results? Must ship today → A; trial with a small circle first → B's trial version (14 days) or C's local loading
If the answers to both questions match, get started right away; if they don't, follow the first one—wherever your users are, that's where your product should be.
🎒 What you'll gain
- A concrete path from "want to learn" to "done": pick a route → follow the lessons → launch your work, with a course to check against at every step
- Know all the hurdles before you start: free quotas, account qualifications, filing and review, store registration fees—you won't get halfway through only to find you're stuck on qualifications
- Your first shareable work: a link (A), a Mini Program people can try by scanning a QR code (B), an extension installed in your own browser (C)
- The practical foundation leading to the advanced handbook: AGENTS.md, the Vercel × AI toolchain, and custom domains — all waiting for you to come back and learn once you have a real project
Ready? The most solid first step into vibe coding practice is A1 Building a website with AI → — the lowest barrier to entry, letting you share the link the same day. If your users are on WeChat, go straight to B1 What is a WeChat Mini Program; if you want to modify your own browser, start with C1 What is a browser extension.