DataCrunch Managed Endpoints API

openapi: 3.2.0
info:
  title: Verda Cloud Public Managed Endpoints API
  description: '# Introduction


    Verda Cloud API allows you to deploy, control instances & more via external code. ## Open API schema


    Download OpenAPI 3.1.0 schema here: /v1/openapi.json


    The API is a RESTful API over HTTPS, all requests & responses use JSON as the data exchange format. ## Official SDKs


    The following SDK packages maintained by Verda Cloud and use this API:


    - **Python SDK**: https://github.com/verda-cloud/sdk-python

    - **Go SDK**: https://github.com/verda-cloud/verdacloud-sdk-go

    - **Terraform** and **OpenTofu** provider: https://github.com/verda-cloud/terraform-provider-verda


    # Quick Start Guide


    We use OAuth 2.0 Client Credentials authentication scheme. You can generate credentials via the console UI https://console.verda.com for external applications to use Verda API. The credentials are used to generate Access & Refresh tokens. All requests to this API must be authenticated using the Access token, except for a few open endpoints. Access tokens have a short lifespan, and can be regenerated using the Refresh token. ## 1. Get Credentials from the Verda Cloud Dashboard


    Go to the Project page => **Keys** => **Cloud API Credentials** section and click on **Create**. These credentials will be used like a username & password for your application. Store the client secret safely - like a password. ## 2. Generate Access Token


    An Access Token is needed to be sent with any API call for authentication. The Access Token has a limited lifetime and is expired after a while as defined in `expires_in`. It is used as a bearer token


    ### To get an access token, call the token endpoint with the credentials obtained in step 1:


    ```json http

    {

    "method": "post",

    "url": "https://api.verda.com/v1/oauth2/token",

    "headers": {

    "Content-Type": "application/json"

    },

    "body": {

    "grant_type": "client_credentials",

    "client_id": "",

    "client_secret": ""

    }

    }

    ```


    A valid response would look similar to this:


    ```json

    {

    "access_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJoZXkiOiJ5b3UgYWN1YWxseSBjaGVja2VkIHRoaXM_In0.0RjcdKQ1NJP9gbRyXITE6LFFLwKGzeeshuubnkkfkb8",

    "token_type": "Bearer",

    "expires_in": 3600,

    "refresh_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ3b3ciOiJhbmQgdGhpcyB0b28_In0.AC5gk-o-MOptUgrouEErlhr8WT3Hg_RR6px6A0I7ZEk",

    "scope": "cloud-api-v1"

    }

    ```


    `access_token` is the token value, to be used in the next step. `expires_in` is the duration in seconds until the token is expired. `refresh_token` can be used to create a new access token if expired. When an Access Token expires, create a new one using the refresh token:


    ```json http

    {

    "method": "post",

    "url": "https://api.verda.com/v1/oauth2/token",

    "headers": {

    "Content-Type": "application/json"

    },

    "body": {

    "grant_type": "refresh_token",

    "refresh_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ3b3ciOiJhbmQgdGhpcyB0b28_In0.AC5gk-o-MOptUgrouEErlhr8WT3Hg_RR6px6A0I7ZEk"

    }

    }

    ```


    The response will include a new token. ## 3. Make REST API Calls


    That''s it, you are ready to use the API, don''t forget to add the Access Token to your calls:


    Let''s call the `/balance` endpoint. Add the Access Token value to the `Authorization` header, preceded by the "Bearer" string and a space:


    ```json http

    {

    "method": "get",

    "url": "https://api.verda.com/v1/balance",

    "headers": {

    "Authorization": "Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJoZXkiOiJ5b3UgYWN1YWxseSBjaGVja2VkIHRoaXM_In0.0RjcdKQ1NJP9gbRyXITE6LFFLwKGzeeshuubnkkfkb8"

    }

    }

    ```


    # Rate Limits


    Authenticated public API requests are rate limited per project in 60 second windows. ## Limits enforced


    - **Project limit**: 500 requests per minute for all authenticated API requests made for the same project

    - **Endpoint limit**: 60 requests per minute for the same project, HTTP method, and request path combination


    The endpoint limit key includes the full request path. Route-specific endpoint policies override the regular endpoint limit for the matching request path.'
  version: v1
  contact:
    name: Verda Cloud Oy
    url: https://verda.com
    email: info@verda.com
servers:
- url: https://api.verda.com
security:
- bearer: []
tags:
- name: Managed Endpoints
paths:
  /v1/managed-endpoints/pricing:
    get:
      operationId: ManagedEndpointsController_getPricing
      parameters:
      - name: currency
        required: false
        in: query
        description: Currency to get the prices in
        schema:
          default: usd
          example: usd
          type: string
          enum:
          - usd
          - eur
      responses:
        '200':
          description: ''
          content:
            application/json:
              schema:
                type: array
                items:
                  $ref: '#/components/schemas/ManagedEndpointPrice'
      summary: Get managed inference endpoint prices
      tags:
      - Managed Endpoints
      security: []
components:
  schemas:
    ManagedEndpointPrice:
      type: object
      properties:
        resource:
          type: string
          description: Managed endpoint / inference model identifier
          example: NVIDIA-H100-80GB-HBM3
        external_id:
          type: string
          description: External managed endpoint / inference model identifier
          example: meta-llama-3-1-8b-instruct
        unit_price:
          type: number
          description: Unit price in the requested currency
          example: 0.99
        unit_name:
          type: string
          enum:
          - generation
          - image
          - video
          - input_token
          - output_token
          - token
          - audio_second
          - video_second
          - inference_second
          - vcpu_second
          - gb_second
          - hour
          - second
          - minute
          - undefined
          - gpu_hour
          - gb_hour
          - gb_month
          - request
          description: Unit the price is charged per
          example: hour
        currency:
          type: string
          enum:
          - usd
          - eur
          description: Currency type
          example: usd
      required:
      - resource
      - external_id
      - unit_price
      - unit_name
      - currency
  securitySchemes:
    bearer:
      scheme: bearer
      bearerFormat: JWT
      type: http
      description: '[Generate the token here](#tag/authentication/POST/v1/oauth2/token). Bearer token to authenticate requests to the public API. You can generate a token by calling the `/v1/oauth2/token` endpoint.'
x-provenance:
  method: harvested
  publisher: DataCrunch
  source: https://api.datacrunch.io/v1/openapi.json
  note: split by refine-openapis from the provider's own published document, verified by fetch (openapi/_original/*.fetch.json);
    stamped by stamp-openapi-provenance.py (roadmap#793)