The company reserves the right to modify terms at its discretion with acceptance deemed upon continued use, lacking a fixed advance notice period for contractual changes. No refund mechanism is provided for material changes during prepaid periods.
Liability is capped at $100 or six months of fees, which limits consumer remedies but is industry-standard for AI. The clause repeatedly uses 'fullest extent permissible' and acknowledges local law overrides.
Courts are preserved without mandatory arbitration, strongly favoring consumer access to justice. However, exclusive jurisdiction in San Francisco and waiver of other jurisdictions' laws creates geographic and legal friction for international or non-California users.
Termination is broadly discretionary and can occur without notice, though paid subscriptions receive pro-rata refunds and an external appeal process is referenced. No explicit right to data export prior to deletion is granted.
Consent is established via browsewrap language rather than explicit clickwrap, which offers weaker consumer protection. The scope is clearly defined to individual services with an appropriate age threshold.
Indemnification is reasonably scoped to user breaches, unlawful conduct, and third-party rights, avoiding blanket coverage for mere service use. It lacks an explicit carve-out for company negligence but does not appear to impose it.
Detailed privacy practices are deferred to a separate policy, reducing transparency within this contract. An opt-out for data usage in model training is provided, and business domain monitoring is disclosed.
Users retain full ownership of inputs and outputs, with a broad but opt-outable license for model training and service improvement. Feedback grants perpetual commercial rights without compensation.
Highly transparent billing with clear auto-renewal disclosure, straightforward online cancellation, 30-day advance notice for price increases, and statutory cooling-off periods for qualifying regions.