Fair-use safeguards

The planned unlimited paid plan is for personal use by a human. It does not permit bots, scripted scraping, automated bulk requests, account sharing, resale or access sold to others. Instant and DeepThink retain their selected behavior; the service does not silently substitute Instant for DeepThink.

Traffic safeguards currently allow up to 30 started answer requests per hour per signed-in account and two simultaneous answers. Public previews allow three requests per hour per client address. A question is limited to 500 characters. Responses have bounded output and processing time. These are service-abuse and availability controls, not a monthly paid-answer quota.

When a safeguard is reached, requests receive an explicit limit response. Wait for current answers to finish or for the stated retry interval. A request that reaches a provider may count even if it fails; automatic model retries are disabled to avoid duplicate usage. Expiring concurrency leases prevent a failed request from locking an account indefinitely.

Payment setup is deferred. Refunds, cancellation details, merchant jurisdiction and final commercial terms require owner review before checkout is enabled. This page is an operational policy draft, not a complete legal agreement.

Answers summarize cited educational source material. They do not provide diagnosis, prescriptions or personalized medical direction.