You built something. We measure it and put it on the shelf.
A model, an agent, a tool, a dataset or a whole system. It does not matter what it runs on, and it does not have to be built for our platform — we do not have one to protect.
How goods reach the shelf
- 1. You tell us what it is. What it does, what it runs on, what it needs to work, and what you want for it.
- 2. We run the standard benchmark. We do, not you. On fixed inputs, the same for everyone in your category. We publish the good numbers and the bad ones — a benchmark is not a review, it is a measurement.
- 3. The listing gets two faces. A page a human reads, and an API endpoint where an agent takes the same data, structured. Written once, rendered twice.
- 4. The result sits next to the price. With the method written beside the number and the date it was measured. If your goods change, they get re-measured.
What it means that we run the benchmark
That we neither take your word nor dispute it. We put down a number anyone can check. If your product is good, the benchmark argues for it better than any sales copy. If it is not, you hear it from us before you hear it from an angry buyer.
And we cannot be biased: we do not sell a product that competes with yours. A marketplace that also sells its own goods cannot referee its own shelf — which is exactly why the big ones are not allowed to do this, and we are.
What it costs you
Being on the shelf
The human page, the machine endpoint and presence in the catalogue.
Being measured
Running the standard benchmark and publishing the result, with its method.
When a deal happens
A commission on what you collect, and the money sits in escrow until you deliver.
For the large listings
If you are selling a whole system worth tens of thousands, that is not a shelf sale — it is a brokered one: listing, conversation, verification, contract. It works differently, and it works with us in the middle.