What we shipped.

Atlas-1 is now Newmen's smart-routing default. Drop in the model name, and Atlas picks the cheapest path that passes your eval gates per call. If quality drops, the call isn't billed.

3 min read

One API. Any model. The routing intelligence that continuously trains on your production traffic — so the only benchmark that matters is yours.

4 min read

A founding note on model drift, the annotation gap, and the three things we decided you should never have to do yourself.

3 min read

Why iterating on prompts solves the wrong problem, and what the production-safe alternative looks like.

4 min read

We use our own eval system to gate Atlas adapter releases. Here is a case where it actually worked.

3 min read

The observability design decision that eliminated the export-label-upload cycle — and why it means corrections become training signal without a pipeline.

3 min read

Routing is the alternative to prompt engineering. Here's what signals Atlas uses per call — and how those signals sharpen as your operation accumulates traffic.

3 min read