scripts/portfolio-demo-beats.md holds the structured 90-second demo runbook and live failure fallback for your full-stack inventory tracker portfolio project.
Run the pre-demo validation command
The command you’ll run to populate this file is `npx demo-prep inventory-tracker –timeout 90s –fallback-assets ./demo-screenshots`, which executes three key checks before generating your structured beat sheet. First, it sends 5 consecutive ping requests to your live Vercel deployment URL to confirm it is accessible, with no active outages, rate limits, or broken redirects. Second, it sends test GET requests to all 4 core API endpoints your demo relies on (dashboard load, inventory fetch, forecast data pull, user auth check) to confirm all return 200 OK statuses and load the pre-seeded sample data you added for demo purposes. Third, it pulls the 6 pre-captured full-resolution screenshots of your core demo flows from your repo’s demo-screenshots directory, to pre-populate the fallback flow if the live site is unreachable. Illustrative example: if your live deployment is rate-limited after 127 consecutive test runs the night before your review, the command will flag the outage immediately instead of you discovering it mid-demo, when you have no time to pivot to a backup plan. The command takes less than 10 seconds to run end to end, so you can execute it as many times as you need as you tweak your project or practice your demo. If the command flags any issues, you can either roll back to the last stable deployment of your project or proceed to use the pre-built fallback flow, with no impact on your demo’s flow or timing.
Save the auto-generated beat sheet output
When you run the command, pipe the output directly to this file with `npx demo-prep inventory-tracker –timeout 90s –fallback-assets ./demo-screenshots > scripts/portfolio-demo-beats.md` to overwrite the blank placeholder with the structured, timed beat sheet. The default output includes a numbered list of 6 tightly timed beats aligned exactly to your 90-second window, with editable placeholders for your specific talking points:
- 0:00–0:10: Opening hook: State your project name, core problem it solves, and the tech stack in one tight, conversational sentence. The pre-filled placeholder reads: “This is Inventory Track, a low-cost stock management tool for small independent coffee shops that cuts manual weekly stock count time by 70%, built with Next.js 14, Supabase, Tailwind CSS, and ShadCN UI.”
- 0:10–0:30: Core feature demo 1: Navigate to the live dashboard, point to the real-time low stock alert widget that triggers in-app and SMS push notifications to shop managers, and click through to a sample low stock oat milk entry to show the auto-generated reorder draft pre-populated with your preferred supplier’s contact information and current pricing.
- 0:30–0:50: Core feature demo 2: Pull up the sales forecasting tab, explain that the lightweight time-series model pulls 90 days of historical sales data to predict stock needs for the next two weeks, and highlight a sample forecast for 2% dairy milk that correctly predicted a 25% sales bump during a local weekend farmers’ market event the previous month.
- 0:50–1:10: Engineering highlight: Call out one specific technical choice you made that demonstrates intentional problem solving, such as the edge function you built to reduce API latency for in-person inventory updates by 62% for shop staff with spotty internet, or the row-level security rules you implemented in Supabase to ensure shop owners can only access their own store’s data, no cross-account data leaks possible.
- 1:10–1:20: Closing value prop: Restate who the tool is built for and one measurable, specific outcome users get when they use it, to leave a clear, memorable takeaway for your reviewer.
- 1:20–1:30: Call to action: Mention that the full repo is public on your GitHub profile, linked directly from your portfolio homepage, and you are happy to walk through any part of the codebase, architecture decisions, or bug fixes in more detail.
Below the beat sheet, the auto-generated output includes a trimmed fallback flow that swaps out steps 2 and 3 for references to your pre-captured screenshots, so you can walk through the exact same demo beats without a live site, with no disruption to your timing. You can edit any of the lines to sound more natural to your speaking style, so the script feels like a guide rather than a rigid text you have to recite word for word.
Commit the finalized script file to your repo
Once you have edited the script to match your preferred talking points and confirmed the fallback flow is accurate, commit the file and the linked fallback screenshots to your project’s main branch. This ensures you have a single source of truth for your demo that you can access from any device, whether you are doing your review from your personal laptop or a borrowed device at an in-person event. Use a clear, specific commit message like “finalize 90s Inventory Track demo script + fallback assets” so you can quickly locate this exact version in your git history if you update the project later and want to generate a new script. Committing the script also means you can reference it easily if you are asked to do a longer demo of the project later, as you can expand on the existing beats instead of building a new runbook from scratch. You do not need to commit every draft of the script, just the final version you plan to use for your portfolio review, to keep your git history clean. If you make small tweaks after practice runs, you can amend the existing commit or push a new small commit with a note about the change, so you always have the latest version saved.
Open this file 24 hours before your portfolio review
This is the next time you will access this file, to do your final pre-review checks and practice runs. First, re-run the pre-demo validation command one more time to confirm your live deployment is still accessible, and cross-reference the newly generated output with your saved script to confirm no recent deployments have broken any of the features you plan to demo. Next, run through the full demo out loud twice, timing yourself to make sure you land within the 85 to 95 second window, adjusting lines if you are consistently running over or under time. Pull up the fallback screenshots to confirm they load correctly, and save a copy of them to your desktop for one-click access if the live site goes down mid-demo, so you do not have to search through your repo folders while your reviewer is waiting. You do not need to memorize the script word for word; the bullet points are designed to be quick cues to keep you on track, so you can speak naturally without rambling or forgetting key details when you feel nervous. If you are doing a virtual review, you can pin a minimized view of this script to the side of your screen off-camera, so you can glance at it as needed without your reviewer noticing. The prep command aligns with the portfolio project guidelines you followed when building the project on Build Next Stack, so you can be confident the script hits all the key points reviewers look for in a portfolio project demo. If you have extra time, you can run through the fallback flow once as well, to make sure you can walk through the screenshots smoothly without the live site, so you feel prepared for any possible outage. Even if you have practiced the demo dozens of times before, this final check ensures you avoid last-minute surprises that can throw off your delivery.