Consent is a product feature.
The choice should be granular, understandable, and visible when it matters. Organization, repository, and session boundaries belong in the workflow. Private should always be a complete path.
Coding agents are becoming part of everyday work. The infrastructure around them should make costs understandable, protect the work by default, and give developers a real say in what happens next.
There is more to a useful coding session than a prompt and an answer. There is the task, the changes, the tests, and the judgment that connects them. We believe that context can be valuable—and that its value begins with the person or team who is entitled to share it.
OptinRouter brings that belief into the workflow: keep an eligible existing inference arrangement and prepare permitted data for a later sale, or accept a funded data-for-discount agreement before qualifying use. Capture and permission to sell remain separate choices.
It is an ambitious product, and this is an early chapter. We are turning the design into something developers can inspect, question, and help shape.
The choice should be granular, understandable, and visible when it matters. Organization, repository, and session boundaries belong in the workflow. Private should always be a complete path.
A usage line should explain what happened and what it cost. A discount should say where it came from. An award should follow a clear standard, with no earnings promises.
Provenance, outcome checks, and quality measurements serve different purposes. We want each claim connected to the evidence that supports it, including what that evidence cannot establish.
Serving traffic and contributed data need distinct systems and permissions. Supplier terms are routing constraints. A feature becomes available when its technical and contractual requirements are met.
The website and workspace demo express the product direction. Live inference, account billing, the local daemon, and contributions are still being developed. We’re inviting the conversations that will sharpen the next release.
Developer, team, model builder, or infrastructure provider—tell us what would make this useful.