The group trip money pot that doesn't start a fight

One shared ledger. Scan receipts. Split unevenly without a spreadsheet. Nudge the person who 'forgot Venmo' without you being the nag.

For Friend groups who still settle up three months later

The problem, sharper

Friend groups who still settle up three months later already tried the generic version and bounced. One shared ledger. Scan receipts. Split unevenly without a spreadsheet. Nudge the person who 'forgot Venmo' without you being the nag. The wedge is the constraint, not another feature list.

Features that make the pitch true

01Capture the job in one screen: One shared ledger. Scan receipts. Split unevenly without a spreadsheet. Nudge the person…
02Audience-aware defaults for friend groups who still settle up three months later.
03Shareable artifact you can send to a teammate in one tap.
04Honest empty states. No fake social proof, no infinite onboarding.
05Export: the user can leave with their data on day one.

This weekend

Friday: landing page, one happy path, seed data.
Saturday: the core loop that makes the pitch true.
Sunday: share link, ugly-but-working export, tell ten people.

Stack

— Next.js + TypeScript
— Local-first persistence, server later
— double-entry basics
— receipt capture
— export for taxes
— invite links

User stories

As a person in this group: Friend groups who still settle up three months later, I can describe the job in one sentence and get a usable first result.
As a skeptic, I can try it without creating an account.
As someone who might build this, I can copy a spec and start this weekend.

Paste into Cursor, Claude Code, or Codex

Build a working product called LumenRoom.

Tagline: The group trip money pot that doesn't start a fight — without the usual product theater.

One-liner: One shared ledger. Scan receipts. Split unevenly without a spreadsheet. Nudge the person who 'forgot Venmo' without you being the nag.

Who it is for: Friend groups who still settle up three months later

Non-goals: do not build a generic finance app. Do not add auth until the core loop works. Do not invent extra personas.

Must-have features:
1. Capture the job in one screen: One shared ledger. Scan receipts. Split unevenly without a spreadsheet. Nudge the person…
2. Audience-aware defaults for friend groups who still settle up three months later.
3. Shareable artifact you can send to a teammate in one tap.
4. Honest empty states. No fake social proof, no infinite onboarding.
5. Export: the user can leave with their data on day one.

Suggested stack:
- Next.js + TypeScript
- Local-first persistence, server later
- double-entry basics
- receipt capture
- export for taxes
- invite links

UX:
- Distinctive, editorial, not purple-gradient SaaS
- Works on a phone
- Empty states teach the product
- One primary action per screen

Ship this weekend:
1. A landing page that states the job in one breath
2. The core loop end to end with sample data
3. A shareable output
4. A README that explains what was intentionally left out

Constraints: TypeScript. Keep the first version local-first if a database would slow you down. Name the product LumenRoom unless you have a clearly better one.