Vision Check API

<b>An API for checking the context and authenticity of visual assets and managing those visionAnalyses</b>

The Compass Vision API provides users with an endpoint to analyze an image by submitting its URL, returning a confidence score indicating whether the image is a deep fake or AI-generated. Upon submitting the image URL, the API processes it using advanced AI detection algorithms to evaluate features commonly associated with manipulated or synthetically generated visuals.

The response will include a confidence score, where a higher score indicates a higher likelihood that the image is a deep fake or AI-generated. This API is ideal for use in content moderation, authenticity verification, and media analysis tools.

You can try these out in your browser with the OpenAPI Spec. Be sure to hit "Authorize 🔓" and enter your credentials first.

Download OpenAPI description

vision.json

vision.yaml

Servers

Mock server

https://docs.blackbird.ai/_mock/vision/

https://api.blackbird.ai/compass/

Create a new vision analysis

Request

Creates a new vision analysis. The resource isn't fully created immediately and you'll want to poll the /visionAnalyses/{id} GET for the result.

Security

clientCredentials or oAuth2Password

Headers

cache-controlstring

Whether to bypass the cached response. Requires permission to bypass the cache.

Value"no-cache"

Bodyapplication/jsonrequired

inputstring(url)required

The input to generate a vision analysis of. Images must have a minumum width and height of 200px and maximum width and height of 4000px. The following file formats are supported - PNG, JPEG, JPG, JPEG 2000, WebP, BMP, GIF, TIFF, TIF, DIB, ICO, EPS, PSD, PPM, PGM, PBM, PNM, TGA, XBM

Example: "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg"

optionsobject(VisionAnalysisOptions)

curl -i -X POST \
  https://docs.blackbird.ai/_mock/vision/visionAnalyses \
  -H 'Authorization: Bearer <YOUR_TOKEN_HERE>' \
  -H 'Content-Type: application/json' \
  -H 'cache-control: no-cache' \
  -d '{
    "input": "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg",
    "options": {
      "explain": true,
      "isVideo": false,
      "needFramesAnnotation": false,
      "framesAnnotationType": [
        "bounding_box",
        "heatmap"
      ]
    }
  }'

Responses

Vision analysis accepted for processing

Bodyapplication/json

idstringrequired

The ID of the vision analysis

Example: "7bf34023-641f-4191-a836-8cd4716266f3"

created_atstring(date-time)required

The datetime the vision analysis was created

Example: "2024-10-10T13:01:29.822Z"

updated_atstring(date-time)required

The datetime the vision analysis was last updated.

Example: "2024-10-10T13:01:29.822Z"

inputstring(url)required

The input to generate a vision analysis of

Example: "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg"

optionsobject(VisionAnalysisOptions)required

options.explainbooleanrequired

Whether to explain the vision analysis result when there is anything worth explaining or not.

Default false

options.isVideoboolean

Whether to use video analysis, requires an additional permission and is rate limited separately

Default false

options.needFramesAnnotationboolean

Whether to annotate the frames of the video with a heatmap or bounding box.

Default false

options.framesAnnotationTypeArray of strings(VisualizationType)

The type of annotation to apply to the frames of the video.

Items Enum"heatmap""bounding_box"

statusstringrequired

The status of the vision analysis generation

Value"processing"

Example: "processing"

Response

application/json

{
  "id": "7bf34023-641f-4191-a836-8cd4716266f3",
  "created_at": "2024-10-10T13:01:29.822Z",
  "updated_at": "2024-10-10T13:01:29.822Z",
  "input": "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg",
  "options": {
    "explain": true,
    "isVideo": false,
    "needFramesAnnotation": true,
    "framesAnnotationType": [ … ]
  },
  "status": "processing"
}

Gets all the vision analyses the caller has submitted.

Request

Note that the result is eventually consistent with creation of a visionAnalysis, so you might not read your own write when you query for all visionAnalyses

Security

clientCredentials or oAuth2Password

Query

startstring(date-time)

Filter vision analyses created after this date (inclusive with ascending sort order, exlusive with descending sort order)

endstring(date-time)

Filter vision analyses created before this date (exclusive with ascending sort order, inclusive with descending sort order)

offsetKeystring

The identifier for the offset to start at. Use the nextOffsetKey from the prior page to get the next page.

sortstring

The direction to sort vision analyses by created_at

Default "desc"

Enum"asc""desc"

curl -i -X GET \
  'https://docs.blackbird.ai/_mock/vision/visionAnalyses?end=2019-08-24T14%3A15%3A22Z&offsetKey=string&sort=asc&start=2019-08-24T14%3A15%3A22Z' \
  -H 'Authorization: Bearer <YOUR_TOKEN_HERE>'

Responses

Vision Analyses found

Bodyapplication/json

dataArray of any(VisionAnalysis)required

data[].idstringrequired

The ID of the vision analysis

Example: "7bf34023-641f-4191-a836-8cd4716266f3"

data[].created_atstring(date-time)required

The datetime the vision analysis was created

Example: "2024-10-10T13:01:29.822Z"

data[].updated_atstring(date-time)required

The datetime the vision analysis was last updated.

Example: "2024-10-10T13:01:29.822Z"

data[].inputstring(url)required

The input to generate a vision analysis of

Example: "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg"

data[].optionsobject(VisionAnalysisOptions)required

data[].options.explainbooleanrequired

Whether to explain the vision analysis result when there is anything worth explaining or not.

Default false

data[].options.isVideoboolean

Whether to use video analysis, requires an additional permission and is rate limited separately

Default false

data[].options.needFramesAnnotationboolean

Whether to annotate the frames of the video with a heatmap or bounding box.

Default false

data[].options.framesAnnotationTypeArray of strings(VisualizationType)

The type of annotation to apply to the frames of the video.

Items Enum"heatmap""bounding_box"

data[].statusstringrequired

The status of the vision analysis generation

Value"processing"

Example: "processing"

Discriminator

nextOffsetKeystring

The offset key of the next page of data.

Response

application/json

{
  "data": [
    { … }
  ],
  "nextOffsetKey": "string"
}

Get input and analysis

Request

Retrieves the vision analysis with the input and the generated analysis around it

Security

clientCredentials or oAuth2Password

Path

idstringrequired

The ID of the vision analysis to retrieve.

Example: 914e5150-7a8c-40a7-b0d9-4738e9d31f9e

curl -i -X GET \
  https://docs.blackbird.ai/_mock/vision/visionAnalyses/914e5150-7a8c-40a7-b0d9-4738e9d31f9e \
  -H 'Authorization: Bearer <YOUR_TOKEN_HERE>'

Responses

Vision Analysis found

Bodyapplication/json

idstringrequired

The ID of the vision analysis

Example: "7bf34023-641f-4191-a836-8cd4716266f3"

created_atstring(date-time)required

The datetime the vision analysis was created

Example: "2024-10-10T13:01:29.822Z"

updated_atstring(date-time)required

The datetime the vision analysis was last updated.

Example: "2024-10-10T13:01:29.822Z"

inputstring(url)required

The input to generate a vision analysis of

Example: "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg"

optionsobject(VisionAnalysisOptions)required

options.explainbooleanrequired

Whether to explain the vision analysis result when there is anything worth explaining or not.

Default false

options.isVideoboolean

Whether to use video analysis, requires an additional permission and is rate limited separately

Default false

options.needFramesAnnotationboolean

Whether to annotate the frames of the video with a heatmap or bounding box.

Default false

options.framesAnnotationTypeArray of strings(VisualizationType)

The type of annotation to apply to the frames of the video.

Items Enum"heatmap""bounding_box"

statusstringrequired

The status of the vision analysis generation

Value"processing"

Example: "processing"

Discriminator

Response

application/json

{
  "id": "7bf34023-641f-4191-a836-8cd4716266f3",
  "created_at": "2024-10-10T13:01:29.822Z",
  "updated_at": "2024-10-10T13:01:29.822Z",
  "input": "https://blackbird.ai/wp-content/uploads/2024/09/GU-BlknW0AA78y0-1.jpeg",
  "options": {
    "explain": true,
    "isVideo": false,
    "needFramesAnnotation": true,
    "framesAnnotationType": [ … ]
  },
  "status": "processing"
}

Vision Check API

Create a new vision analysis

Request

Responses

Was this helpful?

Gets all the vision analyses the caller has submitted.

Request

Responses

Was this helpful?

Get input and analysis

Request

Responses

Was this helpful?

Vision Check API

Create a new vision analysis

RequestExpand all

Responses

Was this helpful?

Gets all the vision analyses the caller has submitted.

Request

Responses

Was this helpful?

Get input and analysis

Request

Responses

Was this helpful?

Request