Terms of Service

Last updated: September 2026. Plain language on purpose; if anything here is unclear, ask us at hello@agenttokens.dev.

1. What this is

AgentTokens ("we") provides metered access to a hosted large language model through an OpenAI‑compatible HTTP API, plus a web dashboard for managing API keys and prepaid credit. By creating an account or using the API you agree to these terms.

2. Accounts and keys

• You sign in with an email address. You are responsible for that inbox and for anything done with keys issued to your account.
• Keep API keys secret. If one leaks, revoke it on the dashboard. Usage made with a leaked key before revocation is still billed.
• One account per person or organisation. Don't create accounts to farm signup credit; we'll close them.

3. Prepaid credit and billing

• Usage is paid from a prepaid balance you top
up via Stripe. We never charge your card automatically.
• Each request is billed at the prices shown on the pricing page at the time of the request, using token counts reported by our gateway. We may change prices; changes apply to future requests only and are announced on the pricing page.
• Credit is non‑refundable, except where a refund is required by applicable law or where we billed you in error. Credit does not expire while your account is open.
• Promotional or signup credit has no cash value and may be removed if we detect abuse.
• When your balance reaches zero, requests are rejected until you top up. A request that was already running may finish and take the balance slightly negative; we won't pursue you for that.

4. Acceptable use

Don't use the service to:

• break the law, or generate content that is illegal where you or we operate;
• attack, probe, or overload the service or other users (this includes deliberately evading fair‑use load shedding);
• resell raw access under your own brand without talking to us first;
• generate spam, malware, or material that sexually exploits minors.

We can suspend or close accounts that violate this section. Unused paid credit on accounts closed for abuse is forfeited.

5. Availability, no SLA

The service is provided "as is". We work hard to keep it up and fast, but there is no uptime guarantee or service‑level agreement unless we've signed a separate written agreement with you. We may change, pause, or discontinue the service; if we shut down for good we'll give at least 30 days' notice and refund unused paid credit.

6. Your content

You keep all rights to your prompts and to the model's outputs. You're responsible for how you use outputs; language models make mistakes and outputs are not professional advice of any kind. We don't use your prompts or outputs to train models. See the Privacy Policy for what we store.

7. Liability

To the extent permitted by law, we're not liable for indirect or consequential losses, lost profits, or lost data arising from use of the service, and our total liability to you is capped at the amount you paid us in the 3 months before the claim.

8. Changes and contact

We may update these terms; material changes are posted here with a new date, and continuing to use the service after that means you accept them. Questions: hello@agenttokens.dev.