AI Supplementary Terms

These AI Supplementary Terms supplement the main Terms of Service. In case of conflict, these AI terms control for AI-related matters.

We do not use your content to train any AI models, neither ours nor our third-party AI providers'. Your content is transmitted to our AI providers solely to process your specific request. We contractually require our AI providers not to use your data for model training.

We currently use:

• Anthropic (Claude) for argument analysis, draft analysis, claim strengthening, consistency checking, and positions map generation.
• Mistral for optical character recognition (OCR) of scanned PDF documents.

Anthropic processes AI requests in the United States. Mistral, an EU-based company, processes OCR requests in the European Union.

You can review each provider's data policies: Anthropic's usage policy and Mistral's terms. We evaluate providers on an ongoing basis and may add or change providers with notice per Section 9 of these terms.

AI features operate on-demand only. Your content is never automatically sent to AI providers. AI processing occurs only when you explicitly invoke an AI feature (such as Strengthen Claim, Analyze Source, Consistency Check, or Build Positions Map). No background AI processing occurs without your action.

When you invoke an AI feature, we send the minimum content necessary to process your request. This may include excerpts from your documents, scaffold structure, or source text, but never your full library. We do not send your account information, email address, or other personal data to AI providers.

Anthropic may retain API inputs and outputs for up to 30 days solely for trust and safety monitoring, after which they are deleted. Mistral retains API data for up to 30 rolling days for abuse monitoring, after which it is automatically deleted. Neither provider uses API data for model training.

• AI features are designed to assist your thinking and writing process, not replace academic judgment.
• AI-generated analysis may contain errors, omissions, or inaccuracies. You are solely responsible for verifying all AI output.
• Argument analysis is assistive, not authoritative. It should not be treated as peer review or editorial guidance.
• warrantd makes no representations regarding the completeness or correctness of AI-generated structural analysis.
• AI output should be independently verified against your sources before relying on it.

You own all AI-generated output produced through your use of the Service, subject to the following:

• AI-generated content may not qualify for copyright protection under current United States law.
• Similar or identical outputs may be generated for other users who provide similar inputs.
• These limitations do not affect your ownership of the underlying documents, scaffolds, and annotations you create.

We may add, replace, or discontinue AI providers or AI features. For material changes (such as adding a new AI provider that will process your content), we will provide at least 30 days' notice via email or in-app notification before the change takes effect. This page will be updated to reflect current providers.

Questions about how warrantd uses AI? Email us at hello@warrantd.com. You can also review our main Terms of Service.