Anthropic API Terms

This document governs consumer use of Claude.ai, Claude Pro, and associated Anthropic individual-facing products, establishing a contract between end users and Anthropic, PBC that explicitly excludes API and commercial console use (governed separately by Commercial Terms of Service). The agreement states that Anthropic may use user inputs and outputs (collectively 'Materials') to train AI models unless users opt out via account settings, though the terms authorize continued use of flagged safety-review content and user-provided feedback for training regardless of opt-out status; the agreement also assigns Anthropic-generated output rights to users 'to the extent permitted by applicable law,' and reserves broad rights to modify, suspend, or terminate services with limited notice obligations. The opt-out carve-outs for safety review and feedback-linked training data represent a notable operational distinction, as does the explicit financial advice prohibition and the automatic renewal structure requiring cancellation at least 24 hours before the renewal date; the agreement asserts broad indemnification obligations from users and limits Anthropic's liability to direct damages not exceeding fees paid in the prior 12 months, which may face scrutiny under EU and UK consumer protection frameworks that restrict liability exclusions for consumer contracts. The terms engage GDPR and UK GDPR for EU and UK users, CCPA for California residents, COPPA for age-gating obligations, and the EU AI Act given deployment of large language model systems; the mandatory arbitration clause with class action waiver, applicable to US users, and the 30-day opt-out window are material compliance considerations, particularly given growing US state-level arbitration scrutiny and California public policy limitations on class action waivers.