Dedicated endpoints
Deploy inference your way.
Bring your own container image for full control over GPU, scaling and runtime, or deploy a model from the Aiter AI World catalog in one step. Both run as dedicated endpoints with reserved GPU and hourly billing.
Custom image
Your image, your rules.
Deploy your own container with full control over GPU selection, runtime and scaling. Ideal for a specific framework, custom model or non-standard setup.
Bring any container image
Deploy from public or private registries. Provide the image, start command and HTTP port.
Full GPU control
Choose the GPU model, GPU count per worker and region.
Flexible scaling
Auto-scale on concurrency or requests per second, or scale manually on a schedule.
Worker logs and monitoring
Per-worker logs, request volume, latency and throughput from the dashboard.
Aiter AI World model
Pick a model. We handle the rest.
Deploy pre-optimised models without configuring images, GPUs or containers. Same dedicated endpoint, simpler setup.
One-step deployment
Select a model and deploy. We handle the infrastructure.
Pre-optimised models
Runtime settings are tuned for inference performance.
Auto-scaling built in
Concurrency or request-based scaling with worker limits.
Managed lifecycle
Deploy, stop and redeploy from the dashboard with metrics and billing.
Need custom or large-scale GPU capacity?
If you cannot find your preferred model, need large capacity or have custom requirements, contact us for a tailored cluster plan.