One endpoint,
every model.
Tare is an OpenAI-compatible gateway in front of every model you use — ours and the ones you bring yourself. One base URL, one key, and every token accounted for.
Overview
Tare is an OpenAI-compatible gateway in front of every model you use. Your application is configured with one base URL and one key; Tare decides which upstream answers.
Two modes, mixable on the same account:
| Platform models | Own channels (BYOK) | |
|---|---|---|
| Token purchaser | Platform | Customer, direct from the provider |
| Platform charge | Platform price list | No per-token charge; routing fee if the contract specifies one |
| Upstream selection | Platform | Customer, on the Routing page |
| Tare provides | One endpoint, failover, usage | The same, plus spend attribution across providers |
[!NOTE] Routing your own providers through Tare does not make the tokens cheaper — you still buy those directly. It puts all your spend in one place, split by model, by API key, and by whatever labels you attach to a call.
Request path
your app ──▶ https://tare.jamerly.ai/v1/chat/completions
│
├─ authenticates your key
├─ selects a route (customer routes first, platform routes as fallback)
├─ forwards, translating the dialect if the upstream is not OpenAI-shaped
└─ records tokens, cost, and attribution
The request body needs no changes. Unrecognised fields are passed through untouched, which is
why cache_control, reasoning, plugins and any parameter the provider adds later keep
working.