Last updated 2026-05-10
These terms govern your use of Dynoyard. By creating an account you agree to them. If you're using Dynoyard for a company, you confirm you have authority to bind the company.
Dynoyard is an app platform for self-hosted AI inference. You select a
blueprint (a model + runtime config), pick a tier, and we provision a
managed OpenAI-compatible endpoint at <name>.dynoyard.app.
Billed monthly through Stripe. Plan changes pro-rate. We give 30 days notice before any price increase. Cancel anytime — your endpoint stays up until the end of the current billing period.
You agree not to use Dynoyard to:
We may suspend or terminate accounts that violate this. We'll give notice where reasonable.
During pilot we operate on best-effort. Target: 99% monthly uptime, <200ms p95 edge latency. We don't yet offer SLA credits — we'll add them with the v1 launch. We'll notify you of major incidents within 24 hours.
You own your prompts, completions, and any data sent to your endpoint. We don't train on it, share it, or retain prompt content. See Privacy Policy.
Each blueprint runs an open-source model under its respective license (Qwen, Llama, DeepSeek). You're responsible for complying with the model's license terms — we'll surface them on each blueprint detail page.
Service provided "as is". To the extent permitted by law, our liability is capped at the fees you paid in the 12 months preceding the claim. We're not liable for indirect, consequential, or special damages.
You can delete your account anytime. We can terminate for material breach of these terms with 30 days notice (immediately for serious abuse). Upon termination we delete your data within 30 days.
Pilot launch jurisdiction: Bahrain (Gulf Cooperation Council). We'll switch to a customer-friendlier venue (likely Delaware) at v1 launch.
Material changes get 30 days email notice. Continued use after that = acceptance.
Questions? [email protected].