Skip to content

Browse Endpoints ​

View endpoints of all projects. The screenshots use vLLM → GPU; the Docker, CPU, and Rebellions menus are the same except for the resource column.

Prerequisites ​

  • An administrator account with access to AI Inference Admin

View all endpoints by type from the Serverless (Docker) and vLLM (GPU, CPU, Rebellions) menus. The four screens are the same except for the resource column.

Sidebar
No.FeatureItemDescription
1-1DockerDocker menu· View Docker endpoints of all projects.
1-2GPU vLLMGPU menu· View GPU vLLM endpoints of all projects.
1-3CPU vLLMCPU menu· View CPU vLLM endpoints of all projects.
1-4Rebellions vLLMRebellions menu· View Rebellions vLLM endpoints of all projects.

Search · Bulk delete ​

Find endpoints across all projects by name and delete selected ones at once. Endpoints cannot be created in the admin app.

Search · Bulk delete
No.FeatureItemDescription
2-1SearchSearch box· Filters the list by endpoint name.
2-2Bulk deleteDelete button· Deletes the endpoints checked in the table. You can delete other users’ endpoints, so take care.

All endpoints ​

Shows endpoints of all projects and clusters in a table.

All endpoints
No.FeatureItemDescription
3-1ColumnsStatusEndpoint status. Green means running, gray paused, and red an error.
NameEndpoint name, with the deployed model (or image) below it.
Project · ClusterThe project the endpoint belongs to and the cluster it runs on. Admin-only columns not shown in the user app.
GPUGPU resources allocated to each replica.
Replicas · Port · Created atMinimum - maximum replicas, request port, and creation date.
3-2Row actionsAction (⋯) button· Opens Service URL, View logs, Send prompt, Pause, Resume, Cancel, Edit, and Delete. They work the same as in the user app.

The row action menu (Service URL, View logs, Send prompt, Pause, Resume, Cancel, Edit, Delete) works the same as in the user app → Manage endpoints.

Other users’ endpoints

In the admin app you can edit and delete endpoints of any project. Check with the owner before stopping or deleting an endpoint in use.

The New Standard for AI Infrastructure