Qubrid AI · Capability

Qubrid AI Compute API

The Qubrid AI Compute API provides programmatic access to GPU cloud infrastructure including NVIDIA H100, H200, and B200 accelerators. Developers can provision and manage GPU instances for AI and machine learning workloads through API calls. The service supports on-demand compute for training, fine-tuning, and batch inference jobs, with usage-based billing and enterprise features such as team collaboration and usage tracking. Instances can be accessed via SSH, Jupyter notebooks, or Visual Studio Code, and support quick-deploy templates for popular frameworks including PyTorch, TensorFlow, Comf

Run with Naftiko QubridAiAPI

What You Can Do

GET
Listinstances — List compute instances
/instances
POST
Createinstance — Create a compute instance
/instances
GET
Getinstance — Retrieve a compute instance
/instances/{instance_id}
DELETE
Deleteinstance — Terminate a compute instance
/instances/{instance_id}
POST
Startinstance — Start a compute instance
/instances/{instance_id}/start
POST
Stopinstance — Stop a compute instance
/instances/{instance_id}/stop
GET
Listgputypes — List available GPU types
/gpus
GET
Listtemplates — List available templates
/templates
GET
Listsshkeys — List SSH keys
/ssh-keys
POST
Createsshkey — Add an SSH key
/ssh-keys
DELETE
Deletesshkey — Remove an SSH key
/ssh-keys/{key_id}

MCP Tools

listinstances

List compute instances

read-only idempotent
createinstance

Create a compute instance

getinstance

Retrieve a compute instance

read-only idempotent
deleteinstance

Terminate a compute instance

idempotent
startinstance

Start a compute instance

stopinstance

Stop a compute instance

listgputypes

List available GPU types

read-only idempotent
listtemplates

List available templates

read-only idempotent
listsshkeys

List SSH keys

read-only idempotent
createsshkey

Add an SSH key

deletesshkey

Remove an SSH key

idempotent

Capability Spec

qubrid-ai-capability.yaml Raw ↑
naftiko: 1.0.0-alpha2
info:
  label: Qubrid AI Compute API
  description: The Qubrid AI Compute API provides programmatic access to GPU cloud infrastructure including NVIDIA H100, H200,
    and B200 accelerators. Developers can provision and manage GPU instances for AI and machine learning workloads through
    API calls. The service supports on-demand compute for training, fine-tuning, and batch inference jobs, with usage-based
    billing and enterprise features such as team collaboration and usage tracking. Instances can be accessed via SSH, Jupyter
    notebooks, or Visual Studio Code, and support quick-deploy templates for popular frameworks including PyTorch, TensorFlow,
    Comf
  tags:
  - Qubrid
  - Ai
  - API
  created: '2026-05-06'
  modified: '2026-05-06'
capability:
  consumes:
  - type: http
    namespace: qubrid-ai
    baseUri: https://platform.qubrid.com/api/v1
    description: Qubrid AI Compute API HTTP API.
    authentication:
      type: bearer
      token: '{{QUBRID_AI_TOKEN}}'
    resources:
    - name: instances
      path: /instances
      operations:
      - name: listinstances
        method: GET
        description: List compute instances
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
      - name: createinstance
        method: POST
        description: Create a compute instance
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: instances-instance-id
      path: /instances/{instance_id}
      operations:
      - name: getinstance
        method: GET
        description: Retrieve a compute instance
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
      - name: deleteinstance
        method: DELETE
        description: Terminate a compute instance
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: instances-instance-id-start
      path: /instances/{instance_id}/start
      operations:
      - name: startinstance
        method: POST
        description: Start a compute instance
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: instances-instance-id-stop
      path: /instances/{instance_id}/stop
      operations:
      - name: stopinstance
        method: POST
        description: Stop a compute instance
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: gpus
      path: /gpus
      operations:
      - name: listgputypes
        method: GET
        description: List available GPU types
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: templates
      path: /templates
      operations:
      - name: listtemplates
        method: GET
        description: List available templates
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: ssh-keys
      path: /ssh-keys
      operations:
      - name: listsshkeys
        method: GET
        description: List SSH keys
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
      - name: createsshkey
        method: POST
        description: Add an SSH key
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: ssh-keys-key-id
      path: /ssh-keys/{key_id}
      operations:
      - name: deletesshkey
        method: DELETE
        description: Remove an SSH key
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
  exposes:
  - type: rest
    port: 8080
    namespace: qubrid-ai-rest
    description: REST adapter for Qubrid AI Compute API.
    resources:
    - path: /instances
      name: listinstances
      operations:
      - method: GET
        name: listinstances
        description: List compute instances
        call: qubrid-ai.listinstances
        outputParameters:
        - type: object
          mapping: $.
    - path: /instances
      name: createinstance
      operations:
      - method: POST
        name: createinstance
        description: Create a compute instance
        call: qubrid-ai.createinstance
        outputParameters:
        - type: object
          mapping: $.
    - path: /instances/{instance_id}
      name: getinstance
      operations:
      - method: GET
        name: getinstance
        description: Retrieve a compute instance
        call: qubrid-ai.getinstance
        outputParameters:
        - type: object
          mapping: $.
    - path: /instances/{instance_id}
      name: deleteinstance
      operations:
      - method: DELETE
        name: deleteinstance
        description: Terminate a compute instance
        call: qubrid-ai.deleteinstance
        outputParameters:
        - type: object
          mapping: $.
    - path: /instances/{instance_id}/start
      name: startinstance
      operations:
      - method: POST
        name: startinstance
        description: Start a compute instance
        call: qubrid-ai.startinstance
        outputParameters:
        - type: object
          mapping: $.
    - path: /instances/{instance_id}/stop
      name: stopinstance
      operations:
      - method: POST
        name: stopinstance
        description: Stop a compute instance
        call: qubrid-ai.stopinstance
        outputParameters:
        - type: object
          mapping: $.
    - path: /gpus
      name: listgputypes
      operations:
      - method: GET
        name: listgputypes
        description: List available GPU types
        call: qubrid-ai.listgputypes
        outputParameters:
        - type: object
          mapping: $.
    - path: /templates
      name: listtemplates
      operations:
      - method: GET
        name: listtemplates
        description: List available templates
        call: qubrid-ai.listtemplates
        outputParameters:
        - type: object
          mapping: $.
    - path: /ssh-keys
      name: listsshkeys
      operations:
      - method: GET
        name: listsshkeys
        description: List SSH keys
        call: qubrid-ai.listsshkeys
        outputParameters:
        - type: object
          mapping: $.
    - path: /ssh-keys
      name: createsshkey
      operations:
      - method: POST
        name: createsshkey
        description: Add an SSH key
        call: qubrid-ai.createsshkey
        outputParameters:
        - type: object
          mapping: $.
    - path: /ssh-keys/{key_id}
      name: deletesshkey
      operations:
      - method: DELETE
        name: deletesshkey
        description: Remove an SSH key
        call: qubrid-ai.deletesshkey
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    port: 9090
    namespace: qubrid-ai-mcp
    transport: http
    description: MCP adapter for Qubrid AI Compute API for AI agent use.
    tools:
    - name: listinstances
      description: List compute instances
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: qubrid-ai.listinstances
      outputParameters:
      - type: object
        mapping: $.
    - name: createinstance
      description: Create a compute instance
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: qubrid-ai.createinstance
      outputParameters:
      - type: object
        mapping: $.
    - name: getinstance
      description: Retrieve a compute instance
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: qubrid-ai.getinstance
      outputParameters:
      - type: object
        mapping: $.
    - name: deleteinstance
      description: Terminate a compute instance
      hints:
        readOnly: false
        destructive: true
        idempotent: true
      call: qubrid-ai.deleteinstance
      outputParameters:
      - type: object
        mapping: $.
    - name: startinstance
      description: Start a compute instance
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: qubrid-ai.startinstance
      outputParameters:
      - type: object
        mapping: $.
    - name: stopinstance
      description: Stop a compute instance
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: qubrid-ai.stopinstance
      outputParameters:
      - type: object
        mapping: $.
    - name: listgputypes
      description: List available GPU types
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: qubrid-ai.listgputypes
      outputParameters:
      - type: object
        mapping: $.
    - name: listtemplates
      description: List available templates
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: qubrid-ai.listtemplates
      outputParameters:
      - type: object
        mapping: $.
    - name: listsshkeys
      description: List SSH keys
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: qubrid-ai.listsshkeys
      outputParameters:
      - type: object
        mapping: $.
    - name: createsshkey
      description: Add an SSH key
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: qubrid-ai.createsshkey
      outputParameters:
      - type: object
        mapping: $.
    - name: deletesshkey
      description: Remove an SSH key
      hints:
        readOnly: false
        destructive: true
        idempotent: true
      call: qubrid-ai.deletesshkey
      outputParameters:
      - type: object
        mapping: $.
binds:
- namespace: env
  keys:
    QUBRID_AI_TOKEN: QUBRID_AI_TOKEN