Terms of Service

1. The service

Caching.ai (the "Service") is a proxy for large-language-model APIs, operated by AI3 Inc. ("we", "us"). You point your SDK at our endpoint with your own provider API keys; we forward your requests, measure cache usage, protect and optionally re-warm your prompt cache, and report savings on your dashboard.

The Service sits between your application and your AI provider. Your contract with each provider (Anthropic, OpenAI, Google, xAI, and others) remains your own — using the Service does not change your obligations to them.

2. Accounts

You need an account to use the Service. You are responsible for the activity that happens under your account and for keeping your credentials and Caching.ai keys confidential. You must provide accurate information and be legally able to enter into this agreement.

3. Fees and billing

Pricing is performance-based: each calendar month we compute your verified savings against provider list prices, subtract the cost of any keep-alive requests we sent on your behalf, and charge 20% of the remaining net savings to your registered payment method after the month closes.

Monthly fees under $5 are waived and never carried over. If no savings are verified, no fee is charged. Fees are exclusive of taxes; where required, taxes are added at the applicable rate.

Savings figures are computed from provider-reported token usage and published list prices. Your dashboard shows the running amount throughout the month.

4. Your responsibilities

You must use the Service only with provider accounts you are authorized to use, and in compliance with each provider's terms and applicable law. You must not use the Service to send unlawful content, to probe or disrupt the Service, or to resell it without our written consent.

You are responsible for the provider API keys you register. You can remove them at any time in the console.

5. Data handling

By default we store token counts, model names, latency, status codes, and hashes of prompt-prefix blocks — not the content of your prompts or responses. If you enable the optional keep-alive feature, we store your prompt prefix (system prompt, tools, and messages up to the last cache breakpoint) encrypted with AES-256-GCM, solely to re-warm your cache; this trade-off is stated on the toggle, and turning it off deletes the stored prefix immediately.

Details are described in our Privacy Policy.

6. Availability and disclaimers

The Service is provided "as is" and "as available". We do not guarantee uninterrupted operation, and savings figures are estimates based on provider-reported usage and list prices. To the maximum extent permitted by law, we disclaim all implied warranties, including merchantability and fitness for a particular purpose.

7. Limitation of liability

To the maximum extent permitted by law, our aggregate liability arising out of or relating to the Service is limited to the fees you paid us in the three months preceding the claim. We are not liable for indirect, incidental, special, or consequential damages, or for loss of profits, data, or goodwill.

8. Suspension and termination

You may stop using the Service and delete your account at any time. We may suspend or terminate accounts that violate these terms or create risk for the Service or other users. Accrued fees remain payable on termination.

9. Changes

We may update these terms as the Service evolves. For material changes we will give notice on the site or by email before they take effect. Continuing to use the Service after the effective date means you accept the updated terms.

10. Governing law and contact

These terms are governed by the laws of the Republic of Korea, without regard to conflict-of-law rules. Questions? Contact us at support@caching.ai.