Skip to content
AppDues
Log in
Contact salesSign up

Shipyard app ideas: scoping a competitive MVP when time is short

A late start is not the same as a lost entry. The original Shipyard write-up made that case with less than six days left before Devpost submissions closed, and its advice still reads as a scoping exercise: pick one creator brief, build the smallest version that satisfies it, and submit through TestFlight or Google Play's testing tracks. This guide reorganizes that material into a decision path — what each brief actually requires, where the build risk sits, and how to avoid losing a finished app to a rules mistake.

Settle the rules before you settle the idea

Shipyard offered seven creator briefs, with one app allowed per brief. The constraints cut both ways: you cannot enter several apps into the same brief, and you cannot enter one app across multiple briefs. Each submission has to map cleanly to a single brief.

That makes brief selection the first real decision, not the last. A workable order for a short window:

  1. List the briefs you could plausibly finish, not the ones that sound most fun.
  2. Write the one sentence a judge would use to describe your app. If it names two briefs, cut features until it names one.
  3. Confirm the build can reach a beta on TestFlight or Google Play's testing tracks. A shipped beta outranks an unfinished polished concept.
  4. If you already have one app ready, treat a second entry as a separate build with its own scope, not a reskin of the first.

Put the included AI tooling to a narrow job

The original post noted that Anthropic's Claude Code Opus 4.6 model was already supported by the tools in the Shipping Container, so participants could use it as part of Shipyard without extra setup. Treat that as the author's account of the event at the time it was written, and confirm current tool availability with the event organizers before relying on it.

The useful habit is narrower than the tool: give the model one transformation to perform instead of asking for "an app". A prompt like "convert this brokerage CSV into a normalized list of holdings with asset type, quantity and cost basis" produces output you can inspect and correct. A prompt like "build an investment tracker" produces a demo you may not be able to finish. Scaffolding and UI iteration are where these tools save the most time; correctness still needs your review.

Brief one: one tracker for every asset class

The VisualFaktory brief starts from a familiar mess: holdings split across stocks, gold, funds, fixed income, real estate and other products, each on a different platform. The requested app logs all of it in one place, refreshes prices where a market feed exists, sets amortization or reminder alerts for products without a listed price, and offers premium risk and diversification views such as country or sector exposure.

The breadth of asset types looks like the hard part, but it is mostly a data-normalization problem, which is where language models are genuinely useful. A realistic MVP accepts spreadsheets, screenshots, PDFs or dictated descriptions, sends them to an API from OpenAI, Anthropic or Google, and stores the result as structured assets the app can chart. Real-time finance data is pulled in only where a feed is available.

For monetization, the source suggested usage-based pricing: AI-assisted imports and advanced risk analysis offered as credit-based features through virtual currencies, so the charge follows the value delivered. That was a suggestion, not a shipped configuration — verify current virtual-currency behavior against product documentation before designing around it.

Brief two: reminders that survive a device switch

The Sam Beckman brief comes from a concrete annoyance: moving between Android and iOS means rebuilding an entire reminder system. The requested app needs custom snoozes straight from notifications, flexible recurring rules, and real sync so that dismissing a reminder on one device clears it everywhere.

For the hackathon, an MVP on either iOS or Android was enough — both platforms were not required. Cross-platform frameworks such as Flutter, Kotlin Multiplatform and React Native fit the brief, and React Native builders like Rork, Blink, Fastshot, Vibecode and Replit make scaffolding and polish fast.

The UI is the easy half. Notifications and cross-device state are the real work, and reliable sync usually means running your own backend. Firebase or Supabase can supply a real-time database, auth and sync without burning the window on infrastructure; the author singled out Supabase's MCP as something an AI agent can drive to set things up. A single-platform reminder app with dependable notifications, clean design and a clear fit to Sam's audience can still stand out.

Brief three: a social app for people living on the road

The Quin Gable brief serves a small, guarded community that is hard to date or befriend in while constantly moving. The requested app combines nomadic dating, activity-based friend finding and a paid builder-help section for van projects, with invite-only or verified access to keep it safe and intentional.

This is a chance to build a social network from scratch. The visible requirements are modest, but accounts, messaging and location-aware discovery still need a backend, and the infrastructure suggestions from the reminders brief apply here as well.

Differentiation decides this brief more than feature count. A single hook that lands with van-lifers beats a longer feature list. One example from the source: a mobile friendship book where you collect connections as you travel from place to place, layered with location-based discovery and messaging. That is a scoped app that feels made for the community rather than a generic social template.

Checklist for choosing between briefs

Run each candidate brief through these questions before writing code:

  • Does it need a backend? Reminders and the van-life social app do; the investment tracker can start with AI APIs plus local storage.
  • Does it need cross-device sync? If yes, budget time for Firebase or Supabase setup before UI polish.
  • Does it imply a paid tier? Decide whether credit-based usage (investment tracker) or a paid section (builder help) matches the value delivered.
  • Can you describe the app in one sentence that names exactly one brief? If not, cut scope.
  • Can you record a three-minute video of the working app? If not, the build is not finished.

Two different ways a submission fails

Late in a hackathon, two failures get mixed together. The first is not shipping: no working beta, no description, no video. The second is shipping something that does not map to one brief, or entering the same app under more than one brief. The first is a build problem you can still fix by cutting scope. The second is a rules problem, and it can waste an app that was finished on time.

Before submitting, confirm the description names a single brief, the video shows the working app rather than slides, and no second entry reuses the same app. The submission itself needs a working app, a clear description and a short three-minute video; the video is the piece most people under-invest in.

Three Shipyard briefs at a glance
BriefCore features requestedMain build riskMonetization angle suggested in the source
Investment tracker for all assets (Josh, VisualFaktory)Log stocks, gold, funds, fixed income, real estate and more; real-time prices where available; amortization/reminder alerts for unlisted products; premium risk and diversification analysisNormalizing many asset types from messy inputsUsage-based credits for AI imports and advanced analysis
Powerful reminders with cross-device sync (Sam Beckman)Custom snoozes from notifications; powerful recurring rules; true sync so a dismissal clears everywhere; iOS and Android in the full vision, one platform for the MVPReliable notification and sync state across devicesNot specified in the source
Social app for van-lifers (Quin Gable)Nomadic dating; activity-based friend finding; paid builder-help section; invite-only or verified accessStanding out from generic social templatesPaid builder-help section

Reflects the briefs as described in the original post at the time it was written. Brief wording, rules and any monetization suggestions are event-specific and should be confirmed with the organizers for any later edition.

Frequently asked questions

Can I submit more than one app to Shipyard?

Yes, but not to the same brief. One app was allowed per brief, each app had to map to exactly one brief, and a single app could not be entered across multiple briefs.

Do I need to build for both iOS and Android for the reminders brief?

No. For the hackathon, an MVP on either iOS or Android was sufficient. The full vision is cross-device sync, but the submission requirement was a single-platform MVP.

What did the submission require?

A working app, a clear description and a short three-minute video showing what you built. The video mattered more than most people expected, so it deserved real time.

How could usage-based pricing fit the investment tracker brief?

The source suggested offering AI-assisted imports and advanced risk analysis as credit-based features using virtual currencies, so the charge follows the value delivered. That was a proposal, not a shipped configuration; confirm current virtual-currency behavior before designing around it.