Orchestration Gateway
Compute and Deployment Management
The Orchestration Gateway manages the lifecycle of model deployments.
Deployments
GET /deployment/deployments
List all active deployments across all compute pools.
org_id
Filter by Organization ID.
POST /deployment/deploy
Create a new model deployment.
model_name
Name of the model to deploy.
image
Docker image to use (e.g., vllm/vllm-openai:latest).
pool_id
Target Compute Pool ID.
replicas
Number of replicas (default: 1).
POST /deployment/terminate
Terminate a running deployment (stops the containers/service).
deployment_id
The ID of the deployment to stop.
DELETE /deployment/delete/{deployment_id}
Permanently remove a stopped or failed deployment record.
deployment_id
The ID of the deployment to delete.
Logs
GET /deployment/logs/{deployment_id}
Retrieve live logs for a specific deployment. Supports DePIN (IPFS) logs.
Model Registry
POST /deployment/registerModel
Register a new model configuration in the system.
model_name
Unique identifier for the model.
backend
Inference backend (e.g., vllm, ollama).
GET /deployment/getModel/{name}/{version}
Get details for a registered model version.