FactQuire — source-verified model pricing/limits feed as a cross-reference for routing

[kimyoohan]

Hi Portkey community,

I wanted to share FactQuire, a source-verified facts feed for commercial LLM APIs: https://factquire.com

Accurate pricing, context windows, and provider limits matter for routing decisions, cost controls, and fallback logic. While auditing community datasets against primary provider sources, we found real drift in both directions — e.g. one widely-used dataset currently lists gemini-2.5-flash-preview-tts at $0.30/M input and $2.50/M output, while Google's pricing page publishes $0.50 (text) input and $10.00 (audio) output. Our own data had an error too (gemini-2.5-flash context window), which we corrected publicly.

FactQuire publishes a machine-readable feed at https://factquire.com/feed.json — each price, limit, and metadata entry includes a verbatim quote from the primary source plus access timestamp, so gateways and routers can validate model metadata without re-auditing provider docs manually.

Feedback from people working on LLM routing would be very welcome.