Akkio Models API

openapi: 3.0.3
info:
  title: Akkio Datasets Models API
  version: v1
  description: 'The Akkio API provides programmatic access to Akkio''s no-code predictive AI platform: manage datasets, train models, and generate predictions. This specification was derived faithfully from Akkio''s first-party client SDKs (github.com/akkio-inc/akkio-python and akkio-node), which document the real paths, methods, request bodies, and authentication. It is a reconstruction, not a provider-published OpenAPI.'
  contact:
    name: Akkio Support
    email: support@akkio.com
    url: https://docs.akkio.com/akkio-help-center
  x-source: https://github.com/akkio-inc/akkio-python
  x-derived: true
servers:
- url: https://api.akk.io/v1
  description: Akkio API v1 production
security:
- apiKeyQuery: []
tags:
- name: Models
  description: Train predictive models and generate predictions.
paths:
  /models:
    get:
      operationId: getModels
      tags:
      - Models
      summary: List models
      description: Return all trained models available to the authenticated account.
      parameters:
      - name: api_key
        in: query
        required: true
        schema:
          type: string
        description: Akkio API key.
      responses:
        '200':
          description: A list of models.
          content:
            application/json:
              schema:
                type: object
                properties:
                  models:
                    type: array
                    items:
                      $ref: '#/components/schemas/Model'
    post:
      operationId: postModel
      tags:
      - Models
      summary: Train a model or make a prediction
      description: 'Overloaded endpoint keyed by request body. With `dataset_id` + `predict_fields`, trains a new model. With `id` + `data`, runs a prediction against an existing model (set `explain: true` for feature-attribution output).'
      requestBody:
        required: true
        content:
          application/json:
            schema:
              oneOf:
              - $ref: '#/components/schemas/CreateModelRequest'
              - $ref: '#/components/schemas/PredictionRequest'
      responses:
        '200':
          description: The trained model or the prediction result.
          content:
            application/json:
              schema:
                oneOf:
                - $ref: '#/components/schemas/Model'
                - $ref: '#/components/schemas/PredictionResult'
    delete:
      operationId: deleteModel
      tags:
      - Models
      summary: Delete a model
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              required:
              - api_key
              - id
              properties:
                api_key:
                  type: string
                id:
                  type: string
                  description: Model id to delete.
      responses:
        '200':
          description: Deletion result.
          content:
            application/json:
              schema:
                type: object
components:
  schemas:
    Model:
      type: object
      properties:
        model_id:
          type: string
        dataset_id:
          type: string
    PredictionRequest:
      type: object
      required:
      - api_key
      - id
      - data
      properties:
        api_key:
          type: string
        id:
          type: string
          description: Model id to predict with.
        data:
          type: array
          items:
            type: object
            additionalProperties: true
        explain:
          type: boolean
          default: false
          description: Return per-feature attribution for each prediction.
    PredictionResult:
      type: object
      additionalProperties: true
    CreateModelRequest:
      type: object
      required:
      - api_key
      - dataset_id
      - predict_fields
      properties:
        api_key:
          type: string
        dataset_id:
          type: string
        predict_fields:
          type: array
          items:
            type: string
          description: Column(s) the model should learn to predict.
        ignore_fields:
          type: array
          items:
            type: string
        extra_attention:
          type: boolean
          default: false
        duration:
          type: integer
          default: 10
          description: Training budget in minutes.
  securitySchemes:
    apiKeyQuery:
      type: apiKey
      in: query
      name: api_key
      description: Akkio API key. Sent as the `api_key` query parameter on GET requests and inside the JSON request body on POST/DELETE requests. Obtain a key at https://app.akk.io/team-settings.