{"command":{"effect":"write","timeoutSeconds":30,"inputSchema":{"type":"object","description":"Update an existing Dedicated Inference. Send a PATCH request to\n`/v2/dedicated-inferences/{dedicated_inference_id}` with updated `spec` and/or\n`access_tokens`. Status will move to updating and return to active when done.\n","properties":{"path":{"type":"object","properties":{"dedicated_inference_id":{"type":"string","format":"uuid","examples":["6b5c619c-359c-44ca-87e2-47e98170c01d"],"description":"A unique identifier for a Dedicated Inference instance."}},"required":["dedicated_inference_id"],"additionalProperties":false},"body":{"$ref":"#/definitions/dedicated_inference_update_request"}},"required":["path","body"],"additionalProperties":false,"definitions":{"dedicated_inference_update_request":{"type":"object","properties":{"spec":{"$ref":"#/definitions/dedicated_inference_spec"},"access_tokens":{"type":"object","description":"Provider tokens for model access (e.g. gated Hugging Face models).","properties":{"hugging_face_token":{"type":"string","description":"Hugging Face token required for gated models.","examples":["$HF_TOKEN"]}},"additionalProperties":false}}},"dedicated_inference_spec":{"type":"object","description":"Structured configuration for a Dedicated Inference deployment.","required":["version","name","region","vpc","enable_public_endpoint","model_deployments"],"properties":{"version":{"type":"integer","description":"Spec version.","examples":[1]},"name":{"type":"string","maxLength":255,"pattern":"^[a-zA-Z0-9_-]+$","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":"Update a Dedicated Inference","id":"digitalocean.dedicatedinferences_patch","example":{"path":{"dedicated_inference_id":"6b5c619c-359c-44ca-87e2-47e98170c01d"},"body":{}},"maxRequests":1,"summary":false}}