FlyRank AI Fluency · Intern Guide
Week 9 · Break it, then launch it · Checkpoint 2

Launch & Keep Building

First you do what beginners skip: try to break your own site on purpose and write down exactly where it cracks. Then you launch for real: your own domain, analytics, the graduate badge, and the habit that keeps the portfolio growing after week ten.

Phase: Submit Estimated hours: 5 Checkpoint 2
This week is a real gate. You do not launch until your must-fixes are handled and the hardening review passes.
The idea this week
  • The part that makes you different from someone who just "made a website": you attack your own site, find where it breaks, and are honest about it. Knowing your limits is the most trusted skill there is.
  • A custom domain is the moment your site stops being "a project I did" and becomes a permanent part of your identity. You plant your flag.
  • Most internships hand you a certificate for a drawer. This hands you a platform. The next project you do becomes the next case here, on the thing you own. It compounds instead of going stale.

Why it matters

Anyone can demo the happy path. The professional difference is knowing exactly where your thing breaks: the empty input, the weird browser, the double submit, and saying so. Then the launch: a custom domain turns "a project" into a permanent part of your online identity, and analytics turns "I hope people visit" into knowing they do. And a portfolio that never gets a second project goes stale; the difference between a class artifact and a career platform is one simple habit, set up now while you still remember how everything works.

Brief

  1. Try to break it: submit your form empty and with garbage, open the site in a browser or device you have not tested, click everything (including demo and repo links), submit twice fast. Note what breaks.
  2. Findability and speed: add basic SEO/meta (page title, description, social-share preview), then search your own name and run a free speed check.
  3. Triage: sort every finding into fix-now and known limitation. Fix the fix-nows, then submit for Checkpoint 2: Hardening Review (mentor or structured peer) with the "where it breaks" list and the fixes. This checkpoint must pass before you launch.
  4. Point a custom domain at your site (a personal domain is a small cost; a clean free subdomain is an acceptable fallback only if budget is truly zero). Add free analytics and confirm the site is fully live over HTTPS on the new address.
  5. Launch hygiene: confirm the social-share preview, favicon, and page titles. Open the final address on your phone once more. Install the FlyRank graduate badge in your footer, linking to your verification page.
  6. Write the plan to keep building: a few lines on exactly where your next case study goes and the steps to add one (reuse the Week 2 three-beat shape), the next real piece of work you intend to add, and a concrete reminder to add it. Keep your Claude Project; it already knows your voice, stack, and identity kit.

Deliverable

The honest "where it breaks" list (fixed versus known-limitation) with evidence the fix-nows are addressed, the live custom-domain URL over HTTPS, a screenshot of analytics installed, the graduate badge visible in your footer, and the "how to add the next case" note with the named next piece and reminder. Post it all in your track thread.

Evaluation criteria (pass/revise)

  • You genuinely tried to break your own site: real edge cases, not just the happy path.
  • Findings are triaged honestly, the fix-nows are actually fixed, and known limitations are named, not hidden.
  • The hardening review happened and its must-fixes are addressed.
  • The site is live on a custom domain (or clean fallback subdomain) over HTTPS, with analytics installed and working.
  • Share preview, favicon, and titles are correct on the real address, and the badge links to your verification page.
  • A concrete plan to keep building exists: a named next piece plus a real reminder, with the Claude Project preserved.
The move this week

Break it, fix it, then plant the flag

Become your own worst user first. Then launch for real, and set up the one habit that keeps the portfolio alive after you graduate.

Attack it, launch it, set the habit. Swap each {highlighted} bit before you send it.

Prompt · Try to break it
Help me break my own portfolio before anyone else does. My live URL: {your_live_url}. Give me a checklist: submit my form empty and with garbage, open the site in a browser or device I have not tested, click everything including demo and repo links, submit twice fast. For each, tell me what a well-behaved site should do, so I can spot where mine misbehaves. Then give me the basic SEO and meta I need before launch (a page title, a description, a social-share preview).
Prompt · Launch on your domain
I am launching my portfolio on a custom domain: {yourname.com}, hosted on {your_host}. Walk me through it, step by step for a beginner: pointing the domain at my site and confirming HTTPS works; confirming the share preview, favicon, and page titles are right on the real address; and adding free, privacy-friendly analytics. After each step, tell me how to confirm it worked. Then give me a final pre-launch checklist to run on my phone.
Prompt · Set your share preview and meta
Help me set the share preview so my link looks intentional when someone pastes it into a chat or a post. My stack: {your_stack}. My proof statement: {paste_your_proof_statement}. Give me the exact page title, meta description, and Open Graph tags (title, description, image) to add, written so they say what I do, plus where to paste them and how to preview the card before I share it.
Prompt · Set up the keep-building habit
I want my portfolio to keep growing after this. Based on how we built it, write me a short "add a new case study" checklist I can follow in 30 minutes whenever I finish a project, reusing my three-beat shape and my existing layout. Then help me name the very next real piece of work I should add: {a_project_you_have_coming_up}, and suggest a concrete reminder to do it.
A worked example

Iris attacks her own site

AI assistant
Iris: The form works, a test message hit my inbox.
AI: Good. Now break it. What happens if you submit it empty?
Iris: It just reloads with no message. Confusing.
AI: Fix-now: add a short "please fill this in" note. What did searching your name show?
Iris: Nothing. My page had no title or description.
AI: Add them. That is the difference between a site that exists and one nobody can find.
Iris: On an old browser my hero font looks plain.
AI: That one is a known limitation. Note it honestly and move on. You do not have to fix everything, you have to know everything.

Iris's contact form works, her site is findable, and she has an honest list of where it is weak. The person who knows their system's limits is the one you trust. Then she pointed irisdesigns.com at her site, turned on analytics, and set a reminder to add her next case the day it ships.

For your track. The launch is the same for everyone. Whatever your path, the keep-building plan is what turns a graded artifact into something that works for you for years.

Watch out for

  • Hiding the cracks. A named limitation is professional. A hidden one is not.
  • Only testing the happy path. The empty form, the weird browser, the double submit. That is where sites break.
  • A vague "I'll add more later." Name the next piece and set a real reminder, or it will not happen.

What good looks like

  • The longest honest "where it breaks" list in your cohort. That is the best work, not the worst.
  • Your name finds your site, and your link previews well.
  • Your site is live on a domain you own, with analytics showing real visitors.
Words you'll hear this week
  • Edge case. The weird situation that breaks things: an empty form, a strange browser.
  • Known limitation. A weak spot you found, chose not to fix for launch, and named honestly.
  • SEO / meta. The basics that help your site be found and shared.
  • Share preview. The card that shows when your link is pasted into a chat or post.
  • Custom domain. Your own web address, like yourname.com.
  • HTTPS. The secure version of a web address, the little padlock. Standard for a real site.
  • Analytics. A free tool that shows you real visitors arriving.
Every term is in the Glossary.

Linked resources

Google Search Central

SEO Starter Guide ↗

The official basics so your site can be found: titles, descriptions, and how search reads a page. Written for beginners.

FreeGuide
 ogp.me

The Open Graph protocol ↗

The exact meta tags that control how your link looks when pasted into a chat or post, straight from the source.

FreeReference
 Netlify Docs

Custom domains & HTTPS ↗

Point your own domain at your site and turn on HTTPS, step by step. The move from a long URL to yourname.com.

FreeDocs
 Cloudflare

Web Analytics ↗

Free, privacy-friendly visitor stats with no cookie banner. See that real people are landing, without spying on them.

FreeTool
 Plausible

Plausible Analytics ↗

A lovely, lightweight, privacy-first analytics tool if you want to go further. Paid after a trial, or free if you self-host.

Paid / trialOpen source