Outcome verification for the commons

Meet the Referee.

PlaceProud is the storefront. The Referee is the infrastructure: the engine that decides whether a real-world job is actually done, so money only moves when the work is real.

Open the live demo How it judges

It scopes the task, grades the proof, gates the money, and sends anything it can't call cleanly to a human. Big tent on what it could verify, hard scope on what ships first.

The pipeline

One photo in. A verdict out.

Four steps turn a single photograph into a clear job, and a finished job into a fair call. The guiding rule is plain: if the Referee can't picture what "done" looks like, it doesn't post the prize.

01

Scope

A photo, a pin, one optional sentence become a clear, checkable job. The safety read happens here, before any money attaches.

02

Post

The prize goes live with its terms set: location, frame, what's in and out, where the bags go, the amount. Fixed from this moment.

03

Prove

The hunter captures in-app: GPS at the pin, framed against the before shot, fresh photos only. A recycled picture doesn't pass.

04

Judge

It grades the before-and-after against the job, with a confidence score. Clean pass, it pays. Anything it's unsure about, a person looks.

ProudTry againAsk a human
Why trust it

The fraud is designed out, not chased after.

The biggest cons can't exist here, because the prize is posted before the payout and the person who posts it is never the person paid to do the work.

The prize comes first

"It was already clean" and "I made the mess myself" both need money posted after the fact. Here it's posted before, by someone who isn't paid for the cleanup.

Capture can't be faked

In-app only, GPS-bound at the pin, framed against the before shot, fresh photos only. A library upload or a borrowed photo gets caught.

A person stands behind the dollar

Anything vague goes to a human, fast, and a hunter wrongly turned down gets a quick second look. The benefit of the doubt costs nothing.

Proven before launch

Tested before a dollar was at stake.

The first fraud it ever saw, a recycled after-photo from a different place, it refused to pay, and named the reason.

The grader is checked against real cleanups and staged cheats before any of it touches money. Early verdicts get audited one by one, so the engine earns the right to decide on its own through a track record, not a promise.

What is really being built

Trash is the first category. Not the point.

Trash is the test case because it's the simplest change a camera can grade: stuff there, then stuff gone. Underneath is a system that can confirm a photographable change to a real place, plus the record of every verdict it has made. Graffiti gone, overgrowth cut back, a tired corner set right... each a future category, added in the order a camera can confirm them. The durable asset is the Referee and its verdict record, and the buyers are anyone who pays for real-world outcomes and needs to know they happened.

Big tent, hard scope. The vision is the whole commons. The first version is litter, on public ground, eighteen and up, nothing in the road or the water. Ambition is wide on purpose; the launch is narrow on purpose; and the vague middle goes to a human, every time.

For cities, districts, and conservancies

Rent the Referee.

If you pay for outcomes in the real world and need to know they actually happened, the verdict record answers it, to the dollar.