PROPERTY · SHARED-COST BILLING
IN PROGRESS

Splitting a shared utility bill without reading anyone's inbox

Bills arrive at a forwarding address, get split per renter, and land in a ledger that survives someone moving out mid-cycle.

CLIENT
Sharehold
ENGAGEMENT
Studio product
STACK
Next.js · TypeScript · Neon Postgres · Better Auth · Resend · LLM extraction

The problem

A landlord with a duplex or a house split into rooms gets one utility bill on a shared meter and is owed a fraction of it by each renter. That means a spreadsheet, percentage maths by hand, an email written per renter, and a mental note of who has paid. That last part rots the moment somebody underpays, overpays, or moves out halfway through a cycle.

The tools that could fix it sit on either side of the gap. Spreadsheets have no audit trail and cannot send anything. Property-management platforms start at portfolio scale and portfolio pricing. Nothing lives in between, which is where every owner-operator with two to four renters actually is.

What we are building

Sharehold turns whatever arrives into an itemised per-renter statement, backed by a ledger that keeps track of who owes what across months. A forwarded email, a PDF attachment, a photo of a paper bill: all the same to it.

Each utility provider gets its own forwarding address belonging to the app, so a bill reaches Sharehold without anyone handing over mailbox access. What arrives is extracted automatically, then shown to the landlord for review before it counts as a real charge: a bad extraction can never quietly become money someone owes. Renters carry a percentage share per category, so the split is computed rather than remembered, and payments record as exact, over or under, with the difference carried forward instead of fudged.

The hard part

This is a ground-up rebuild of a personal tool, and the thing being rebuilt is the part that worked. The original, Next Bill Manager, read the landlord's Gmail directly to find bills and spot e-transfer payments. It was the fastest way to make it work for one person, and the single reason it could never be published: it excludes anyone not on Gmail, asks strangers for read access to their entire mailbox over a forty-dollar water bill, and puts Google's restricted-scope security assessment on the critical path.

Replacing mailbox access with app-owned forwarding addresses removes all three at once, and means the product asks for no Google permissions at all.

The other stubborn case is a Canadian one. Several providers send "your bill is ready, log in to view" with no amount anywhere in the message, which is exactly what breaks every automation these landlords have already tried. So manual entry is a first-class path rather than a fallback, and it has to be finishable in under thirty seconds.

Status

In active development, and marked in progress rather than dressed up as delivered. It will appear here with real numbers once it has run billing cycles for people who are not us.