Terms of service

The short version: pay for the tokens you use, keep your keys secret, do not use the service to break the law, and we serve you as reliably as we can without promising perfection.

Effective 2026-08-16 · contact [email protected]

The service

engy provides an OpenAI-compatible API for running open-weight language models, billed per token. Most models run on GPU infrastructure we operate; some are served by independent operators on the engy network, which changes how your data is handled and is set out under Data below. Using the API or the dashboard means you accept these terms. If you use engy on behalf of an organization, you accept them for that organization.

Accounts and keys

You are responsible for all usage under your API keys. Keep them secret; rotate or revoke a key immediately if you suspect it leaked. We may suspend keys or accounts that violate these terms or generate abusive traffic.

Billing

Service is prepaid: usage draws down your credit balance at the per-token rates published on /pricing. Prices can change; the posted price at request time is what you pay. Purchased credits are refundable on request, minus usage. Promotional and granted credits are not refundable and may expire.

Acceptable use

Do not use the service for unlawful activity, to infringe the rights of others, or to attack, probe, or overload the service itself. You are responsible for complying with the license terms of the models you invoke.

Outputs

You own your prompts and, to the extent we hold any rights in them, we assign model outputs to you. Model outputs are generated by open-weight models and can be wrong; you are responsible for reviewing them before relying on them.

Availability

We aim for high availability but offer no uptime guarantee at this stage. At capacity the API returns early 429 responses rather than queueing. Models may be added, upgraded, or retired; material changes are announced on the site.

Data

Handling of prompts, outputs, and account data is described in the privacy policy. Summary: zero retention of content, metadata only for billing.

Zero data retention (ZDR) applies to models served entirely on hardware engy operates. Those models are marked ZDR on /pricing. For them we do not store your prompts or model outputs, and never train on them.

qwen3.6-35b-a3b is excluded from ZDR. Requests to that model are routed across public, permissionless miners on the engy network: independent operators running hardware without confidential computing, who hold administrative access to their own machines. Operators accept a no-retention policy as a condition of serving traffic, and engy audits their responses for integrity, but on hardware engy does not control, non-retention is a policy and contractual commitment rather than something engy can technically guarantee. Do not send data to qwen3.6-35b-a3b that you are unwilling to expose to a third-party operator. Every other model on /pricing is served only by engy and is covered by ZDR.

Liability

The service is provided as is, without warranties of any kind. To the maximum extent permitted by law, our total liability for any claim arising from the service is limited to the amount you paid us in the three months before the claim. We are not liable for indirect or consequential damages, or for what you build with model outputs.

Termination

You can stop using the service at any time; email us to close your account. We may suspend or terminate service for breach of these terms, with a refund of purchased credits when the termination is not for abuse.

Changes

We may update these terms; material changes are posted here with a new effective date. Continued use after a change means acceptance.