{"command":{"effect":"write","timeoutSeconds":30,"inputSchema":{"type":"object","description":"Create a new Dedicated Inference for your team. Send a POST request to\n`/v2/dedicated-inferences` with a `spec` object (version, name, region, vpc,\nenable_public_endpoint, model_deployments) and optional `access_tokens` (e.g.\nhugging_face_token for gated models). The response code 202 Accepted indicates\nthe request was accepted for processing; it does not indicate success or failure.\nThe token value is returned only on create; store it securely.\n","properties":{"body":{"$ref":"#/definitions/dedicated_inference_create_request"}},"required":["body"],"additionalProperties":false,"definitions":{"dedicated_inference_create_request":{"type":"object","required":["spec"],"properties":{"spec":{"$ref":"#/definitions/dedicated_inference_spec"},"access_tokens":{"type":"object","additionalProperties":{"type":"string"},"description":"Key-value pairs for provider tokens (e.g. Hugging Face).","examples":[{"hugging_face_token":"$HF_TOKEN"}]}}},"dedicated_inference_spec":{"type":"object","description":"Structured configuration for a Dedicated Inference deployment.","required":["version","name","region","vpc","enable_public_endpoint","model_deployments"],"properties":{"version":{"type":"integer","description":"Spec version.","examples":[1]},"name":{"type":"string","maxLength":255,"pattern":"^[a-zA-Z0-9_-]+$","description":"Name of the Dedicated Inference. Must be unique within the team.","examples":["new-dedicated-inference"]},"region":{"type":"string","enum":["atl1","nyc2","tor1"],"description":"DigitalOcean region where the Dedicated Inference is hosted.","examples":["atl1"]},"vpc":{"type":"object","required":["uuid"],"properties":{"uuid":{"type":"string","format":"uuid","description":"VPC UUID for the Dedicated Inference.","examples":["997615ce-132d-4bae-9270-9ee21b395e5d"]}}},"enable_public_endpoint":{"type":"boolean","description":"Whether to expose a public LLM endpoint."},"model_deployments":{"type":"array","minItems":1,"items":{"$ref":"#/definitions/model_deployment_spec"},"description":"At least one model deployment is required."}}},"model_deployment_spec":{"type":"object","description":"Configuration for a single model deployment.","properties":{"model_id":{"type":"string","description":"Used to identify an existing deployment when updating; empty means create new.","examples":[""]},"model_slug":{"type":"string","description":"Model identifier (e.g. Hugging Face slug).","examples":["mistral/mistral-7b-instruct-v3"]},"model_provider":{"type":"string","enum":["hugging_face"],"description":"Model provider.","examples":["hugging_face"]},"workload_config":{"type":"object","description":"Workload-specific configuration (e.g. ISL/OSL in future)."},"accelerators":{"type":"array","items":{"$ref":"#/definitions/accelerator_config_spec"},"description":"Accelerator configuration for this deployment."}}},"accelerator_config_spec":{"type":"object","required":["scale","type","accelerator_slug"],"properties":{"scale":{"type":"integer","minimum":1,"description":"Number of accelerator instances.","examples":[1]},"type":{"type":"string","description":"Accelerator type (e.g. prefill_decode).","examples":["prefill_decode"]},"accelerator_slug":{"type":"string","description":"DigitalOcean GPU slug.","examples":["gpu-mi300x1-192gb"]}}}}},"name":"Create a Dedicated Inference","id":"digitalocean.dedicatedinferences_create","example":{"body":{"spec":{"version":1,"name":"new-dedicated-inference","region":"atl1","vpc":{"uuid":"997615ce-132d-4bae-9270-9ee21b395e5d"},"enable_public_endpoint":false,"model_deployments":[{}]}}},"maxRequests":1,"summary":false}}