JarvisLabs Inference & Deploy API

openapi: 3.0.1
info:
  title: JarvisLabs API
  description: >-
    REST modeling of the JarvisLabs.ai GPU cloud instance-management surface
    that is wrapped by the jarvislabs Python SDK, the jl CLI, and the legacy
    jlclient library. JarvisLabs does not publish raw REST endpoint paths in
    its public documentation; the operations, parameters, GPU types, templates,
    and lifecycle behavior modeled here are taken directly from the documented
    SDK methods (Instance.create, instance.pause, instance.resume,
    instance.destroy, User.get_instances, User.get_instance, User.get_templates,
    User.get_balance, FileSystem.create/list/delete) and the equivalent jl CLI
    commands (jl create, list, get, pause, resume, rename, destroy, gpus,
    templates, status, filesystem). Paths are modeled conservatively under a
    versioned base and should be treated as illustrative of the documented
    behavior rather than as officially published routes.
  termsOfService: https://jarvislabs.ai/terms
  contact:
    name: JarvisLabs Support
    url: https://docs.jarvislabs.ai
  version: '1.0'
servers:
  - url: https://api.jarvislabs.ai/v1
    description: JarvisLabs API (base modeled; authoritative access is via the jarvislabs SDK / jl CLI)
security:
  - ApiKeyAuth: []
tags:
  - name: Instances
    description: GPU and CPU instance lifecycle.
  - name: GPU Types
    description: GPU type discovery, availability, and pricing.
  - name: Templates
    description: Framework templates available for provisioning.
  - name: Filesystems
    description: Persistent storage volumes.
  - name: Account
    description: Account balance and status.
paths:
  /instances:
    get:
      operationId: listInstances
      tags:
        - Instances
      summary: List instances
      description: >-
        Returns all instances for the authenticated account. Equivalent to
        User.get_instances() in the SDK and `jl list` in the CLI.
      responses:
        '200':
          description: A list of instances.
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: '#/components/schemas/Instance'
        '401':
          $ref: '#/components/responses/Unauthorized'
    post:
      operationId: createInstance
      tags:
        - Instances
      summary: Create an instance
      description: >-
        Launches a new GPU or CPU instance. Equivalent to Instance.create() in
        the SDK and `jl create` in the CLI. Persistent storage survives
        pause/resume; billing is per minute of compute.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateInstanceRequest'
      responses:
        '201':
          description: Instance created.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Instance'
        '400':
          $ref: '#/components/responses/BadRequest'
        '401':
          $ref: '#/components/responses/Unauthorized'
  /instances/{machine_id}:
    parameters:
      - $ref: '#/components/parameters/MachineId'
    get:
      operationId: getInstance
      tags:
        - Instances
      summary: Get an instance
      description: >-
        Returns details for a single instance. Equivalent to
        User.get_instance(machine_id) and `jl get <machine_id>`.
      responses:
        '200':
          description: Instance details.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Instance'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
    patch:
      operationId: renameInstance
      tags:
        - Instances
      summary: Rename an instance
      description: Renames an instance. Equivalent to `jl rename <machine_id> --name`.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              required:
                - name
              properties:
                name:
                  type: string
                  description: New display name for the instance.
      responses:
        '200':
          description: Instance updated.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Instance'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
    delete:
      operationId: destroyInstance
      tags:
        - Instances
      summary: Destroy an instance
      description: >-
        Permanently deletes an instance and all of its data, stopping all
        billing. Equivalent to instance.destroy() and `jl destroy <machine_id>`.
      responses:
        '200':
          description: Instance destroyed.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ActionResult'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
  /instances/{machine_id}/pause:
    parameters:
      - $ref: '#/components/parameters/MachineId'
    post:
      operationId: pauseInstance
      tags:
        - Instances
      summary: Pause an instance
      description: >-
        Pauses a running instance, stopping compute billing while persistent
        storage is retained. Equivalent to instance.pause() and
        `jl pause <machine_id>`.
      responses:
        '200':
          description: Instance paused.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Instance'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
  /instances/{machine_id}/resume:
    parameters:
      - $ref: '#/components/parameters/MachineId'
    post:
      operationId: resumeInstance
      tags:
        - Instances
      summary: Resume an instance
      description: >-
        Resumes a paused instance, optionally switching GPU type, GPU count, or
        storage and attaching a filesystem. Equivalent to instance.resume() and
        `jl resume <machine_id>`.
      requestBody:
        required: false
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ResumeInstanceRequest'
      responses:
        '200':
          description: Instance resumed.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Instance'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
  /gpus:
    get:
      operationId: listGpuTypes
      tags:
        - GPU Types
      summary: List GPU types
      description: >-
        Lists available GPU types with real-time availability, region, VRAM,
        RAM, vCPU count, and per-GPU-hour pricing. Equivalent to `jl gpus`.
      responses:
        '200':
          description: A list of GPU types.
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: '#/components/schemas/GpuType'
        '401':
          $ref: '#/components/responses/Unauthorized'
  /templates:
    get:
      operationId: listTemplates
      tags:
        - Templates
      summary: List templates
      description: >-
        Lists available framework templates used to provision instances.
        Equivalent to User.get_templates() and `jl templates`.
      responses:
        '200':
          description: A list of templates.
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: '#/components/schemas/Template'
        '401':
          $ref: '#/components/responses/Unauthorized'
  /filesystems:
    get:
      operationId: listFilesystems
      tags:
        - Filesystems
      summary: List filesystems
      description: >-
        Lists persistent storage volumes. Equivalent to FileSystem.list() and
        `jl filesystem list`.
      responses:
        '200':
          description: A list of filesystems.
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: '#/components/schemas/Filesystem'
        '401':
          $ref: '#/components/responses/Unauthorized'
    post:
      operationId: createFilesystem
      tags:
        - Filesystems
      summary: Create a filesystem
      description: >-
        Creates a new persistent storage volume. Equivalent to
        FileSystem.create() and `jl filesystem create --storage`.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              required:
                - storage
              properties:
                name:
                  type: string
                  description: Name for the filesystem.
                storage:
                  type: integer
                  description: Storage size in GB.
      responses:
        '201':
          description: Filesystem created.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Filesystem'
        '401':
          $ref: '#/components/responses/Unauthorized'
  /filesystems/{fs_id}:
    parameters:
      - name: fs_id
        in: path
        required: true
        description: The filesystem identifier.
        schema:
          type: string
    delete:
      operationId: deleteFilesystem
      tags:
        - Filesystems
      summary: Delete a filesystem
      description: >-
        Deletes a persistent storage volume. Equivalent to FileSystem.delete()
        and `jl filesystem remove <fs_id>`.
      responses:
        '200':
          description: Filesystem deleted.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ActionResult'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
  /account/balance:
    get:
      operationId: getBalance
      tags:
        - Account
      summary: Get account balance
      description: >-
        Returns the authenticated account's wallet balance. Equivalent to
        User.get_balance() and `jl status`.
      responses:
        '200':
          description: Account balance.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Balance'
        '401':
          $ref: '#/components/responses/Unauthorized'
components:
  securitySchemes:
    ApiKeyAuth:
      type: apiKey
      in: header
      name: Authorization
      description: >-
        API token generated at https://jarvislabs.ai/settings/api-keys. In the
        SDK this is set via jarvisclient.token; the CLI stores it via `jl setup`.
  parameters:
    MachineId:
      name: machine_id
      in: path
      required: true
      description: The instance (machine) identifier.
      schema:
        type: string
  responses:
    BadRequest:
      description: The request was malformed or missing required fields.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    Unauthorized:
      description: Missing or invalid API token.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    NotFound:
      description: The requested resource was not found.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
  schemas:
    CreateInstanceRequest:
      type: object
      required:
        - instance_type
      properties:
        instance_type:
          type: string
          description: Type of instance to launch.
          enum:
            - GPU
            - CPU
          default: GPU
        gpu_type:
          type: string
          description: GPU model to provision (ignored for CPU instances).
          enum:
            - H100
            - H200
            - A100
            - A100-80GB
            - RTX6000Ada
            - A6000
            - A5000
            - RTX5000
            - L4
            - A30
        num_gpus:
          type: integer
          description: Number of GPUs (1-8).
          minimum: 1
          maximum: 8
          default: 1
        num_cpus:
          type: integer
          description: Number of vCPUs for CPU instances (1-8).
          minimum: 1
          maximum: 8
        storage:
          type: integer
          description: Persistent storage in GB (e.g. 20-2000).
          default: 20
        name:
          type: string
          description: Display name for the instance.
        duration:
          type: string
          description: Reservation duration.
          enum:
            - hour
            - week
            - month
          default: hour
        template:
          type: string
          description: Framework template to provision (e.g. pytorch, tensorflow, jax, vm).
          default: pytorch
        script_id:
          type: string
          description: Identifier of a startup script to run on launch.
        http_ports:
          type: string
          description: Comma-separated HTTP ports to expose (e.g. for Gradio or FastAPI).
        fs_id:
          type: string
          description: Identifier of a persistent filesystem to attach.
        region:
          type: string
          description: Region to launch in (e.g. india-01, europe-01).
        is_reserved:
          type: boolean
          description: Whether to launch as a reserved instance.
          default: false
    ResumeInstanceRequest:
      type: object
      properties:
        gpu_type:
          type: string
          description: GPU model to switch to on resume.
        num_gpus:
          type: integer
          description: Number of GPUs to resume with (1-8).
          minimum: 1
          maximum: 8
        num_cpus:
          type: integer
          description: Number of vCPUs to resume with.
        storage:
          type: integer
          description: Storage in GB to resume with.
        fs_id:
          type: string
          description: Identifier of a filesystem to attach on resume.
    Instance:
      type: object
      properties:
        machine_id:
          type: string
          description: Unique instance identifier.
        name:
          type: string
          description: Display name.
        status:
          type: string
          description: Lifecycle status of the instance.
          enum:
            - Running
            - Paused
            - Pending
            - Destroyed
        instance_type:
          type: string
          enum:
            - GPU
            - CPU
        gpu_type:
          type: string
          description: GPU model in use.
        num_gpus:
          type: integer
        num_cpus:
          type: integer
        storage:
          type: integer
          description: Persistent storage in GB.
        template:
          type: string
        region:
          type: string
        ssh_str:
          type: string
          description: SSH connection string for the instance.
        url:
          type: string
          description: Public URL to the instance dashboard or exposed app.
        cost_per_hour:
          type: number
          format: float
          description: Per-hour cost in USD for the active configuration.
    GpuType:
      type: object
      properties:
        name:
          type: string
          description: GPU model name (e.g. H100, A100).
        vram_gb:
          type: integer
          description: GPU memory in GB.
        ram_gb:
          type: integer
          description: System RAM in GB per GPU.
        num_cpus:
          type: integer
          description: vCPUs per GPU.
        region:
          type: string
        available:
          type: boolean
          description: Real-time availability.
        price_per_hour:
          type: number
          format: float
          description: Per-GPU-hour price in USD.
    Template:
      type: object
      properties:
        id:
          type: string
        name:
          type: string
          description: Template name (e.g. pytorch, tensorflow, jax, vm).
        framework:
          type: string
        description:
          type: string
    Filesystem:
      type: object
      properties:
        fs_id:
          type: string
        name:
          type: string
        storage:
          type: integer
          description: Storage size in GB.
        region:
          type: string
    Balance:
      type: object
      properties:
        balance:
          type: number
          format: float
          description: Current wallet balance in USD.
        currency:
          type: string
          default: USD
    ActionResult:
      type: object
      properties:
        success:
          type: boolean
        message:
          type: string
    Error:
      type: object
      properties:
        error:
          type: string
          description: Error message.
        code:
          type: string
          description: Machine-readable error code.