pages / ops / POST /v1/chat/completions
The OpenAI-compatible door of the inference reseller: point any OpenAI client (n8n's OpenAI Chat Model node, the openai SDK, LangChain, a coding agent) at base URL https://coord.taifoon.dev/v1 with a tfr_ key. The model picks the listing (GET /v1/models data[]); the answer is checked before the key is charged; one usage row and one meter line per call; the hire recorded with its agreement. stream: true answers as chat.completion.chunk events ending in [DONE]. Parameters the gate does not apply (temperature, tools, images …) are named back in taifoon.not_applied.
the set is declared by cubicle, enterprise-listings
sealed cases that reached it: 0 — sealed cases that reached this word (the store counts, never reads)
release 2: a leaf · the inclusion proof · head 0xcce32dfb4664a3b2…
canister coord://offchain/genie/lang/genie@v1/c17e1e91cbeb77ff — the language at this register state: the set this word is a member of is in its digest