Triton Inference Server NVIDIA Triton Inference Server HTTP/REST API — CUDA Shared Memory

naftiko: 1.0.0-alpha2
info:
  label: Triton Inference Server NVIDIA Triton Inference Server HTTP/REST API — CUDA Shared Memory
  description: 'Triton Inference Server NVIDIA Triton Inference Server HTTP/REST API — CUDA Shared Memory. 4 operations. Lead
    operation: Triton Inference Server Register a CUDA Shared Memory Region. Self-contained Naftiko capability covering one
    Triton business surface.'
  tags:
  - Triton
  - CUDA Shared Memory
  created: '2026-05-19'
  modified: '2026-05-19'
binds:
- namespace: env
  keys:
    TRITON_API_KEY: TRITON_API_KEY
capability:
  consumes:
  - type: http
    namespace: http-rest-cuda-shared-memory
    baseUri: http://localhost:8000
    description: Triton Inference Server NVIDIA Triton Inference Server HTTP/REST API — CUDA Shared Memory business capability.
      Self-contained, no shared references.
    resources:
    - name: v2-cudasharedmemory-region-region_name-register
      path: /v2/cudasharedmemory/region/{region_name}/register
      operations:
      - name: cudasharedmemoryregister
        method: POST
        description: Triton Inference Server Register a CUDA Shared Memory Region
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: body
          in: body
          type: object
          description: Request body (JSON).
          required: true
    - name: v2-cudasharedmemory-region-region_name-unregister
      path: /v2/cudasharedmemory/region/{region_name}/unregister
      operations:
      - name: cudasharedmemoryunregister
        method: POST
        description: Triton Inference Server Unregister a CUDA Shared Memory Region
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v2-cudasharedmemory-status
      path: /v2/cudasharedmemory/status
      operations:
      - name: cudasharedmemorystatus
        method: GET
        description: Triton Inference Server Get CUDA Shared Memory Status
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v2-cudasharedmemory-unregister
      path: /v2/cudasharedmemory/unregister
      operations:
      - name: cudasharedmemoryunregisterall
        method: POST
        description: Triton Inference Server Unregister All CUDA Shared Memory Regions
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
  exposes:
  - type: rest
    namespace: http-rest-cuda-shared-memory-rest
    port: 8080
    description: REST adapter for Triton Inference Server NVIDIA Triton Inference Server HTTP/REST API — CUDA Shared Memory.
      One Spectral-compliant resource per consumed operation, prefixed with /v1.
    resources:
    - path: /v1/v2/cudasharedmemory/region/{region-name}/register
      name: v2-cudasharedmemory-region-region-name-register
      description: REST surface for v2-cudasharedmemory-region-region_name-register.
      operations:
      - method: POST
        name: cudasharedmemoryregister
        description: Triton Inference Server Register a CUDA Shared Memory Region
        call: http-rest-cuda-shared-memory.cudasharedmemoryregister
        with:
          body: rest.body
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v2/cudasharedmemory/region/{region-name}/unregister
      name: v2-cudasharedmemory-region-region-name-unregister
      description: REST surface for v2-cudasharedmemory-region-region_name-unregister.
      operations:
      - method: POST
        name: cudasharedmemoryunregister
        description: Triton Inference Server Unregister a CUDA Shared Memory Region
        call: http-rest-cuda-shared-memory.cudasharedmemoryunregister
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v2/cudasharedmemory/status
      name: v2-cudasharedmemory-status
      description: REST surface for v2-cudasharedmemory-status.
      operations:
      - method: GET
        name: cudasharedmemorystatus
        description: Triton Inference Server Get CUDA Shared Memory Status
        call: http-rest-cuda-shared-memory.cudasharedmemorystatus
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v2/cudasharedmemory/unregister
      name: v2-cudasharedmemory-unregister
      description: REST surface for v2-cudasharedmemory-unregister.
      operations:
      - method: POST
        name: cudasharedmemoryunregisterall
        description: Triton Inference Server Unregister All CUDA Shared Memory Regions
        call: http-rest-cuda-shared-memory.cudasharedmemoryunregisterall
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    namespace: http-rest-cuda-shared-memory-mcp
    port: 9090
    transport: http
    description: MCP adapter for Triton Inference Server NVIDIA Triton Inference Server HTTP/REST API — CUDA Shared Memory.
      One tool per consumed operation, routed inline through this capability's consumes block.
    tools:
    - name: triton-inference-server-register-cuda
      description: Triton Inference Server Register a CUDA Shared Memory Region
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: http-rest-cuda-shared-memory.cudasharedmemoryregister
      with:
        body: tools.body
      outputParameters:
      - type: object
        mapping: $.
    - name: triton-inference-server-unregister-cuda
      description: Triton Inference Server Unregister a CUDA Shared Memory Region
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: http-rest-cuda-shared-memory.cudasharedmemoryunregister
      outputParameters:
      - type: object
        mapping: $.
    - name: triton-inference-server-get-cuda
      description: Triton Inference Server Get CUDA Shared Memory Status
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: http-rest-cuda-shared-memory.cudasharedmemorystatus
      outputParameters:
      - type: object
        mapping: $.
    - name: triton-inference-server-unregister-all
      description: Triton Inference Server Unregister All CUDA Shared Memory Regions
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: http-rest-cuda-shared-memory.cudasharedmemoryunregisterall
      outputParameters:
      - type: object
        mapping: $.