Committed-use pricing, a dedicated region, and the controls your security team will ask for.
Billed through Stripe, drawn down by the tokens you use (input and output) plus a flat $0.0001 per request. Live usage shows in your console.
The same unit your LLM uses. Input covers your queries and stored memories; output is the context WOS returns.
No. WOS only meters its own retrieval tokens. With BYOK, your model provider bills you directly for theirs.
Recall returns a bounded, fixed-size context, so output is capped. Set a hard spend cap and usage stops at your limit.
Yes. Opt in per search with cache_control and repeated or extended queries reuse the previous result at 10% of the normal rate. The first request pays a write premium (2x for a 5-minute TTL, 3x for 1 hour), and any write to the store invalidates its cache instantly. Details in the docs.
No subscription and no seats. You pay only for the tokens you use, on the same rate from prototype to production. Balance is prepaid, topped up from $5.
Add a card, grab an API key, top up at least $5, and you're metered from the first call.
UNDER CONSTRUCTION
Top floor or bottom, we're hauling bricks and climbing ladders, setting it one solid block at a time. We'll be back with something better, soon.