Annotate Item

POSThttps://api.confident-ai.com/v2/annotation-queues/{annotationQueueId}/items/{queueItemId}/annotate

Records your team's annotation of one queue item and marks it complete, returning the ids of what was written. Send annotations for criteria ratings, formResponses for answers to the queue's annotation form, or both; answering the form requires annotatorEmail. Send markAsCompleted: false to leave the item in the pending list.

POST/v2/annotation-queues/{annotationQueueId}/items/{queueItemId}/annotate

curl -X POST "https://api.confident-ai.com/v2/annotation-queues/{annotationQueueId}/items/{queueItemId}/annotate" \
  -H "CONFIDENT_API_KEY: <PROJECT-API-KEY>" \
  -H "Content-Type: application/json" \
  -d '{
  "annotations": [
    {
      "rating": 1,
      "type": "FIVE_STAR_RATING",
      "name": "Helpfulness",
      "explanation": "Answered the question and cited the right source.",
      "expectedOutput": "Mount Everest is 8,848 metres tall.",
      "expectedOutcome": "The user learns how tall Mount Everest is.",
      "imagesMapping": {}
    }
  ],
  "formResponses": [
    {
      "label": "How helpful was the answer?",
      "value": "Very helpful"
    }
  ],
  "annotatorEmail": "jane@acme.com",
  "flagged": false,
  "markAsCompleted": true
}'

200

{
  "success": true,
  "data": {
    "annotationIds": [
      "<ANNOTATION-ID>"
    ],
    "formResponseIds": [
      "<ANNOTATION-FORM-RESPONSE-ID>"
    ]
  },
  "link": "https://app.confident-ai.com/project/<PROJECT-ID>/annotation-queues/<ANNOTATION-QUEUE-ID>",
  "deprecated": false
}

Headers

CONFIDENT_API_KEYstringRequired
The API key of your Confident AI project.

Path parameters

annotationQueueIdstringRequired
The id of the annotation queue the item belongs to.
queueItemIdstringRequired
The id of the queue item.

Request body

annotationslist of objects
The criteria ratings to record on the item, one entry per criterion.
Show 7 propertiesHide 7 properties
- ratingintegerRequired
  The rating to record: 0 or 1 for a THUMBS_RATING, 1 to 5 for a FIVE_STAR_RATING.
- typeenum
  This is the type of annotation, which can be either thumbs rating or five star rating.
  Show 2 enum valuesHide 2 enum values
  - FIVE_STAR_RATING
  - THUMBS_RATING
- namestring
  The criterion this rating is for, matching a custom annotation option in this project. Omit it to rate the built-in criterion.
- explanationstring
  Why the rating was given.
- expectedOutputstring
  The output the target should have produced. Only for an item holding a trace or span.
- expectedOutcomestring
  The outcome the conversation should have reached. Only for an item holding a thread.
- imagesMappingobject
  Images referenced by [DEEPEVAL:IMAGE:<key>] markers in the text fields, keyed by that marker's key.
formResponseslist of objects
The answers to the fields of the queue's annotation form. Sending them requires annotatorEmail.
Show 2 propertiesHide 2 properties
- labelstringRequired
  The label of the form field being answered, exactly as the form spells it.
- valueany
  The answer, in the shape the field's type expects: a string for TEXT, a number for NUMBER or FLOAT, a boolean for BOOLEAN, one of selectOptions for SELECT, and a list of them for MULTI_SELECT.
annotatorEmailstring
The email address of the project member the work is credited to. Required when formResponses are sent, and what makes the annotation visible on the platform.
flaggedboolean
Whether to flag the item for a second opinion.
markAsCompletedboolean
Whether to mark the item annotated, taking it out of the pending list. Defaults to true.

Response

Annotate Item succeeded.

successboolean
Indicates if the request was successful.
dataobject
What one item's annotation wrote.
Show 2 propertiesHide 2 properties
- annotationIdslist of strings
  The ids of the annotations recorded, one per criterion rated.
- formResponseIdslist of strings
  The ids of the form answers recorded, one per field answered.
linkstring
This is the URL of the resource on the Confident AI platform.
deprecatedboolean
Indicates if this endpoint is deprecated.

Built byConfident AI