openapi: 3.2.0
info:
  title: Eden Ai Question Answer API
  version: '2.0'
  description: 'Operations tagged Question Answer across 2 of this provider''s published API definitions: eden-ai-image-features-openapi.json, eden-ai-video-features-openapi.json. Each path carries the servers of the definition it was published in.'
servers:
- url: https://api.edenai.run/v2
tags:
- name: Question Answer
paths:
  /image/question_answer/:
    post:
      operationId: image_question_answer_create
      description: 'Available Providers


        |Provider|Model|Version|Price|Billing unit|

        |----|----|-------|-----|------------|

        |**openai**|**gpt-4-turbo**|`v1`|30.0 (per 1000000 token)|1 token

        |**openai**|**gpt-4o-2024-08-06**|`v1`|10.0 (per 1000000 token)|1 token

        |**openai**|**gpt-4.1-2025-04-14**|`v1`|8.0 (per 1000000 token)|1 token

        |**openai**|**gpt-5**|`v1`|10.0 (per 1000000 token)|1 token

        |**openai**|**o1**|`v1`|60.0 (per 1000000 token)|1 token

        |**openai**|**o1-mini**|`v1`|12.0 (per 1000000 token)|1 token

        |**openai**|**gpt-4o**|`v1`|10.0 (per 1000000 token)|1 token

        |**openai**|**gpt-4o-mini**|`v1`|0.6 (per 1000000 token)|1 token

        |**google**|**gemini-1.5-pro**|`v1`|10.0 (per 1000000 token)|1 token

        |**google**|**gemini-2.5-flash**|`v1`|2.5 (per 1000000 token)|1 token

        |**google**|**gemini-2.5-pro**|`v1`|15.0 (per 1000000 token)|1 token

        |**google**|**gemini-2.0-flash**|`v1`|0.4 (per 1000000 token)|1 token

        |**google**|**gemini-1.5-flash**|`v1`|0.6 (per 1000000 token)|1 token


        Supported Languages


        |Name|Value|

        |----|-----|

        |**English**|`en`|

        |**French**|`fr`|

        |**German**|`de`|

        |**Italian**|`it`|

        |**Spanish**|`es`|


        Supported Models


        Default Models


        |Name|Value|

        |----|-----|

        |**openai**|`gpt-4o`|

        |**google**|`gemini-2.5-pro`|'
      summary: Question Answer
      tags:
      - Question Answer
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/imagequestion_answerQuestionAnswerRequest'
            examples:
              RequestExample:
                value:
                  providers: openai,google
                  question: What are the logos on the image ?
                  temperature: 0
                  max_tokens: 64
                  file_url: http://edenai-resource-example.jpg
                summary: Request Example
          multipart/form-data:
            schema:
              $ref: '#/components/schemas/imagequestion_answerQuestionAnswerRequest'
            examples:
              RequestExample:
                value:
                  providers: openai,google
                  file: /edenai/edenai/features/image/samples/data/logo_detection.jpeg
                  question: What are the logos on the image ?
                  temperature: 0
                  max_tokens: 64
                summary: Request Example
        required: true
      security:
      - FeatureApiAuth: []
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/imagequestion_answerResponseModel'
              examples:
                ResponseExample:
                  value:
                    openai:
                      answers:
                      - Logos - Facebook, Apple, Amazon, Netflix, Google
                      usage:
                        completion_tokens: 19
                        prompt_tokens: 1659
                        total_tokens: 1678
                        completion_tokens_details:
                          accepted_prediction_tokens: 0
                          audio_tokens: 0
                          reasoning_tokens: 0
                          rejected_prediction_tokens: 0
                        prompt_tokens_details:
                          audio_tokens: 0
                          cached_tokens: 0
                      cost: 0.0
                    google:
                      answers:
                      - Logo - Facebook, Apple, Amazon, Netflix, and Google
                      usage:
                        completion_tokens: 17
                        prompt_tokens: 770
                        total_tokens: 787
                        completion_tokens_details: null
                        prompt_tokens_details: null
                      cost: 0.0
                  summary: Response Example
          description: ''
        '400':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/BadRequest'
          description: ''
        '500':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
          description: ''
        '403':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
          description: ''
        '404':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/NotFoundResponse'
          description: ''
    servers:
    - url: https://api.edenai.run/v2
  /video/question_answer/:
    post:
      operationId: video_question_answer_create
      description: 'Available Providers


        |Provider|Model|Version|Price|Billing unit|

        |----|----|-------|-----|------------|

        |**google**|**gemini-3-flash-preview**|`v1Beta`|3e-06 (per 1 token)|1 token

        |**google**|**gemini-3-pro-preview**|`v1Beta`|1.2e-05 (per 1 token)|1 token

        |**google**|**gemini-2.0-flash**|`v1Beta`|0.4 (per 1000000 token)|1 token

        |**google**|-|`v1Beta`|0.6 (per 1000000 token)|1 token

        |**google**|**gemini-2.0-flash-lite-preview**|`v1Beta`|0.3 (per 1000000 token)|1 token

        |**google**|**gemini-3.1-pro-preview**|`v1Beta`|1.2e-05 (per 1 token)|1 token

        |**google**|**gemini-2.5-flash-lite**|`v1Beta`|0.4 (per 1000000 token)|1 token

        |**google**|**gemini-2.5-pro**|`v1Beta`|15.0 (per 1000000 token)|1 token

        |**google**|**gemini-2.5-flash**|`v1Beta`|2.5 (per 1000000 token)|1 token

        |**google**|**gemini-3.1-flash-lite-preview**|`v1Beta`|1.5e-06 (per 1 token)|1 token

        |**google**|**gemini-2.5-pro-preview-03-25**|`v1Beta`|15.0 (per 1000000 token)|1 token

        |**google**|**gemini-2.0-flash-lite**|`v1Beta`|0.3 (per 1000000 token)|1 token

        |**amazon**|-|`boto3 (v1.29.6)`|3.2 (per 1000000 token)|1 token

        |**amazon**|**amazon.nova-pro-v1:0**|`boto3 (v1.29.6)`|3.2 (per 1000000 token)|1 token


        Supported Models


        Default Models


        |Name|Value|

        |----|-----|

        |**google**|`gemini-2.5-flash`|

        |**amazon**|`amazon.nova-pro-v1:0`|'
      summary: Question Answer
      tags:
      - Question Answer
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/videoquestion_answerQuestionAnswerRequest'
            examples:
              RequestExample:
                value:
                  providers: amazon,google
                  text: describe this video
                  max_tokens: 6500
                  temperature: 1
                  file_url: http://edenai-resource-example.mp4
                summary: Request Example
          multipart/form-data:
            schema:
              $ref: '#/components/schemas/videoquestion_answerQuestionAnswerRequest'
            examples:
              RequestExample:
                value:
                  providers: amazon,google
                  file: /edenai/edenai/features/video/samples/data/faces.mp4
                  text: describe this video
                  max_tokens: 6500
                  temperature: 1
                summary: Request Example
        required: true
      security:
      - FeatureApiAuth: []
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/videoquestion_answerResponseModel'
              examples:
                ResponseExample:
                  value:
                    amazon:
                      answer: The video starts with a person holding a smartphone and showing a picture of a room with a table, a chair, and a potted plant. The table has a glass and a skull-shaped object on it. The room has a checkered floor and a window with a curtain. The person then shows a picture of a building with a tower and a statue on top. The building has a clock tower and a tower with a statue of a person on top. The building is surrounded by trees and has a fence around it. The person then shows a picture of a man sitting in front of a desk with two monitors, a keyboard, and a mouse. He is holding a bottle and a pen in his hands. He is looking at the monitor on the right. The desk has a keyboard, a mouse, and a bottle. Behind him is a chair. The room is illuminated by a neon light.
                      usage:
                        completion_tokens: 180
                        prompt_tokens: 3461
                        total_tokens: 3641
                      cost: 0.0
                    google:
                      answer: 'Here is a description of the video.


                        The video is a tutorial on how to create AI hyperlapses. The video opens with a shot of a phone displaying a short video of a table with a glass of water and a small skull on it. The narrator then explains that AI hyperlapses are currently going viral on YouTube and Instagram. The video then shows several examples of AI hyperlapses, including a statue of a woman and a lion, a church dome, and a statue of a woman and child. The narrator then shows how to create these videos using a phone and computer.'
                      finish_reason: STOP
                      usage: null
                      cost: 0.0
                  summary: Response Example
          description: ''
        '400':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/BadRequest'
          description: ''
        '500':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
          description: ''
        '403':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
          description: ''
        '404':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/NotFoundResponse'
          description: ''
    servers:
    - url: https://api.edenai.run/v2
components:
  schemas:
    StatusEnum:
      enum:
      - sucess
      - fail
      type: string
    FieldError:
      type: object
      properties:
        <parameter_name>:
          type: array
          items:
            type: string
      required:
      - <parameter_name>
    NotFoundResponse:
      type: object
      properties:
        details:
          type: string
          default: Not Found
    NestedError:
      type: object
      properties:
        type:
          type: string
        message:
          type: string
      required:
      - message
      - type
    Error:
      type: object
      properties:
        error:
          $ref: '#/components/schemas/NestedError'
      required:
      - error
    imagequestion_answerQuestionAnswerDataClass:
      properties:
        answers:
          items:
            type: string
          title: Answers
          type: array
        original_response:
          default: null
          description: original response sent by the provider, hidden by default, show it by passing the `show_original_response` field to `true` in your request
          title: Original Response
        status:
          allOf:
          - $ref: '#/components/schemas/StatusEnum'
          title: Status
      required:
      - status
      title: imagequestion_answerQuestionAnswerDataClass
      type: object
    imagequestion_answerQuestionAnswerRequest:
      type: object
      properties:
        settings:
          type: string
          default: {}
          description: "A dictionnary or a json object to specify specific models to use for some providers. <br>                     It can be in the following format: {\"google\" : \"google_model\", \"ibm\": \"ibm_model\"...}.\n                     "
        providers:
          type: array
          items:
            type: string
            minLength: 1
          description: 'It can be one (ex: **''amazon''** or **''google''**) or multiple provider(s) (ex: **''amazon,microsoft,google''**)             that the data will be redirected to in order to get the processed results.<br>             Providers can also be invoked with specific models (ex: providers: **''amazon/model1, amazon/model2, google/model3''**)'
        fallback_providers:
          type: array
          items:
            type: string
          default: []
          description: "Providers in this list will be used as fallback if the call to provider in `providers` parameter fails.\n    To use this feature, you must input **only one** provider in the `providers` parameter. but you can put up to 5 fallbacks.\n\nThey will be tried in the same order they are input, and it will stop to the first provider who doesn't fail.\n\n\n*Doesn't work with async subfeatures.*\n    "
          maxItems: 5
        response_as_dict:
          type: boolean
          default: true
          description: "Optional : When set to **true** (default), the response is an object of responses with providers names as keys : <br> \n                  ``` {\"google\" : { \"status\": \"success\", ... }, } ``` <br>\n                When set to **false** the response structure is a list of response objects : <br> \n                   ``` [{\"status\": \"success\", \"provider\": \"google\" ... }, ] ```. <br>\n                  "
        attributes_as_list:
          type: boolean
          default: false
          description: "Optional : When set to **false** (default) the structure of the extracted items is list of objects having different attributes : <br>\n     ```{'items': [{\"attribute_1\": \"x1\",\"attribute_2\": \"y2\"}, ... ]}``` <br>\n     When it is set to **true**, the response contains an object with each attribute as a list : <br>\n     ```{ \"attribute_1\": [\"x1\",\"x2\", ...], \"attribute_2\": [y1, y2, ...]}``` "
        show_base_64:
          type: boolean
          default: true
        show_original_response:
          type: boolean
          default: false
          description: "Optional : Shows the original response of the provider.<br>\n        When set to **true**, a new attribute *original_response* will appear in the response object."
        file:
          type: string
          format: binary
          description: 'File to analyse in binary format to be used with *content-type*: **multipart/form-data** <br> **Does not work with application/json !**'
          pattern: (?:jpg|jpeg|png|tiff)$
        file_url:
          type:
          - string
          - 'null'
          format: uri
          description: 'File **URL** to analyse to be used with with *content-type*: **application/json**.'
        temperature:
          type: number
          format: double
          maximum: 1
          minimum: 0
          default: 0.0
          description: Higher values mean the model will take more risks and value 0 (argmax sampling) works better for scenarios with a well-defined answer.
        question:
          type: string
          minLength: 1
          description: Question about the image
        max_tokens:
          type: integer
          maximum: 2048
          minimum: 1
          default: 1000
          description: The maximum number of tokens to generate in the completion. The token count of your prompt plus max_tokens cannot exceed the model's context length.
      required:
      - providers
    NestedBadRequest:
      type: object
      properties:
        type:
          type: string
        message:
          $ref: '#/components/schemas/FieldError'
      required:
      - message
      - type
    BadRequest:
      type: object
      properties:
        error:
          $ref: '#/components/schemas/NestedBadRequest'
      required:
      - error
    imagequestion_answerResponseModel:
      properties:
        alephalpha:
          $ref: '#/components/schemas/imagequestion_answerQuestionAnswerDataClass'
          default: null
        openai:
          $ref: '#/components/schemas/imagequestion_answerQuestionAnswerDataClass'
          default: null
        google:
          $ref: '#/components/schemas/imagequestion_answerQuestionAnswerDataClass'
          default: null
      title: imagequestion_answerResponseModel
      type: object
    videoquestion_answerQuestionAnswerRequest:
      type: object
      properties:
        settings:
          type: string
          default: {}
          description: "A dictionnary or a json object to specify specific models to use for some providers. <br>                     It can be in the following format: {\"google\" : \"google_model\", \"ibm\": \"ibm_model\"...}.\n                     "
        providers:
          type: array
          items:
            type: string
            minLength: 1
          description: 'It can be one (ex: **''amazon''** or **''google''**) or multiple provider(s) (ex: **''amazon,microsoft,google''**)             that the data will be redirected to in order to get the processed results.<br>             Providers can also be invoked with specific models (ex: providers: **''amazon/model1, amazon/model2, google/model3''**)'
        fallback_providers:
          type: array
          items:
            type: string
          default: []
          description: "Providers in this list will be used as fallback if the call to provider in `providers` parameter fails.\n    To use this feature, you must input **only one** provider in the `providers` parameter. but you can put up to 5 fallbacks.\n\nThey will be tried in the same order they are input, and it will stop to the first provider who doesn't fail.\n\n\n*Doesn't work with async subfeatures.*\n    "
          maxItems: 5
        response_as_dict:
          type: boolean
          default: true
          description: "Optional : When set to **true** (default), the response is an object of responses with providers names as keys : <br> \n                  ``` {\"google\" : { \"status\": \"success\", ... }, } ``` <br>\n                When set to **false** the response structure is a list of response objects : <br> \n                   ``` [{\"status\": \"success\", \"provider\": \"google\" ... }, ] ```. <br>\n                  "
        attributes_as_list:
          type: boolean
          default: false
          description: "Optional : When set to **false** (default) the structure of the extracted items is list of objects having different attributes : <br>\n     ```{'items': [{\"attribute_1\": \"x1\",\"attribute_2\": \"y2\"}, ... ]}``` <br>\n     When it is set to **true**, the response contains an object with each attribute as a list : <br>\n     ```{ \"attribute_1\": [\"x1\",\"x2\", ...], \"attribute_2\": [y1, y2, ...]}``` "
        show_base_64:
          type: boolean
          default: true
        show_original_response:
          type: boolean
          default: false
          description: "Optional : Shows the original response of the provider.<br>\n        When set to **true**, a new attribute *original_response* will appear in the response object."
        file:
          type: string
          format: binary
          description: 'File to analyse in binary format to be used with *content-type*: **multipart/form-data** <br> **Does not work with application/json !**'
          pattern: (?:mp4|mpeg|mov|avi|x\-flx|mpg|webm|wmv|3gpp)$
        file_url:
          type:
          - string
          - 'null'
          format: uri
          description: 'File **URL** to analyse to be used with with *content-type*: **application/json**.'
        temperature:
          type: number
          format: double
          maximum: 1
          minimum: 0
          default: 0.0
          description: Higher values mean the model will take more risks and value 0 (argmax sampling) works better for scenarios with a well-defined answer.
        max_tokens:
          type: integer
          maximum: 3000000
          minimum: 1
          description: The maximum number of tokens to generate in the completion. This value, combined with the token count of your prompt, cannot exceed the model's context length.
        text:
          type: string
          minLength: 1
          description: Question about the video
      required:
      - providers
      - text
    videoquestion_answerQuestionAnswerDataClass:
      properties:
        answer:
          title: Answer
          type: string
        finish_reason:
          default: null
          title: Finish Reason
          type: string
        original_response:
          default: null
          description: original response sent by the provider, hidden by default, show it by passing the `show_original_response` field to `true` in your request
          title: Original Response
        status:
          allOf:
          - $ref: '#/components/schemas/StatusEnum'
          title: Status
      required:
      - answer
      - status
      title: videoquestion_answerQuestionAnswerDataClass
      type: object
    videoquestion_answerResponseModel:
      properties:
        amazon:
          $ref: '#/components/schemas/videoquestion_answerQuestionAnswerDataClass'
          default: null
        google:
          $ref: '#/components/schemas/videoquestion_answerQuestionAnswerDataClass'
          default: null
      title: videoquestion_answerResponseModel
      type: object
  securitySchemes:
    FeatureApiAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
x-refined-from:
- eden-ai-image-features-openapi.json
- eden-ai-video-features-openapi.json