On-prem nodes

Run inference on your own GPUs behind the same API.

Register a node

Organization owners and admins register nodes in the console. A node must be reachable over the organization's private Tailscale mesh. Nodes are isolated per organization: other organizations cannot call them.

Call it

curl https://api.eggsplain.com/v1/chat/completions \
  -H "Authorization: Bearer $EGGSPLAIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "onprem/<slug>",
  "messages": [
    { "role": "user", "content": "Hello from my own hardware." }
  ]
}'

Nest-synced deployments appear in /v1/models under their model display name. Their previous onprem/<slug> IDs continue to work.

Requests to an onprem/ model never fall back to cloud providers, so your data stays on your hardware.

Vision

Add the :vision flag to a model ref for multimodal requests, for example sovereign/egg-mom-1:vision.