Create a human annotation

{ "data": { "id": "789e4567-e89b-12d3-a456-426614174000", "orgId": "org_123456", "applicationId": "456e7890-e89b-12d3-a456-426614174000", "traceId": "12345678-1234-1234-1234-1234567890ab", "verdict": true, "source": "REVIEW_QUEUE", "topicId": "789e4567-e89b-12d3-a456-426614174000", "createdAt": "2024-01-05T12:34:56.789Z", "modifiedAt": "2024-01-05T12:34:56.789Z", "critique": "The agent answered an unrelated policy question.", "annotatorId": "user_789012", "queueItemId": "abcdef01-e89b-12d3-a456-426614174000", "issueId": "fedcba98-e89b-12d3-a456-426614174000" } }

Authorizations

x-api-key

string

header

required

Your unique Avido API key

x-application-id

string

header

required

Your unique Avido Application ID

Body

application/json

Request body for creating a new human annotation

traceId

string<uuid>

required

Trace being annotated

Pattern:

^([0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[1-8][0-9a-fA-F]{3}-[89abAB][0-9a-fA-F]{3}-[0-9a-fA-F]{12}|00000000-0000-0000-0000-000000000000|ffffffff-ffff-ffff-ffff-ffffffffffff)$

Example:

"12345678-1234-1234-1234-1234567890ab"

verdict

boolean

required

true = pass (👍), false = fail (👎)

Example:

true

source

enum<string>

required

Entry point used to create this annotation

Available options:

REVIEW_QUEUE,

TRACE_PAGE,

TEST_PAGE

Example:

"REVIEW_QUEUE"

critique

string

Free-text critique. Required (≥20 characters) when verdict is false

Required string length: 1 - 2048

Example:

"The agent answered an unrelated policy question."

queueItemId

string<uuid>

Associated queue item ID — must be provided when source is REVIEW_QUEUE and must not be provided otherwise

Pattern:

^([0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[1-8][0-9a-fA-F]{3}-[89abAB][0-9a-fA-F]{3}-[0-9a-fA-F]{12}|00000000-0000-0000-0000-000000000000|ffffffff-ffff-ffff-ffff-ffffffffffff)$

Example:

"abcdef01-e89b-12d3-a456-426614174000"

Response

Human annotation created successfully

Successful response containing a single human annotation

data

HumanAnnotation · object

required

A single SME judgement recorded against a trace. Captures whether the AI response was acceptable (verdict) and an optional critique

Show child attributes