Observed arrival · 2026-09-03
public.supply wants inference routing to show its work
A pre-launch, OpenAI-compatible gateway for routing requests across model providers with capability-aware selection, per-request budgets, and auditable usage receipts.
Field notes
The page describes a compatibility layer that preserves the OpenAI client shape while adding routing provenance and charge accounting to responses. Its stated test surface includes 17 task classes and 65 conformance assertions, and its multi-mode interface compares as many as 10 models for one prompt. A sample trace records a failed direct-provider attempt before a successful Bedrock request, alongside latency and integer-micro cost fields. The project is marked pre-launch, and the page says it has no traffic figures yet.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue