Skip to main content
POST
/
v1
/
transcript-evaluation
Create transcript evaluation
curl --request POST \
  --url https://analytics-api.voiceflow.com/v1/transcript-evaluation \
  --header 'Content-Type: application/json' \
  --header 'authorization: <api-key>' \
  --data '
{
  "projectID": "<string>",
  "name": "<string>",
  "enabled": true,
  "prompt": "<string>",
  "type": "boolean",
  "truePrompt": "<string>",
  "falsePrompt": "<string>",
  "description": null,
  "settings": null
}
'
{
  "evaluation": {
    "id": "<string>",
    "projectID": "<string>",
    "name": "<string>",
    "description": "<string>",
    "default": true,
    "enabled": true,
    "averageCost": 123,
    "prompt": "<string>",
    "settings": {
      "model": "gpt-3.5-turbo-1106",
      "realtime": {
        "voice": "<string>",
        "eagerness": "<string>"
      },
      "maxTokens": 123,
      "temperature": 123,
      "reasoningEffort": "minimal"
    },
    "systemTag": "<string>",
    "type": "boolean",
    "truePrompt": "<string>",
    "falsePrompt": "<string>"
  }
}

Authorizations

authorization
string
header
required

Voiceflow Dialog Manager API key (VF.DM) or Workspace API key (VF.WS)

Body

application/json

Create a binary evaluation.

projectID
string
required

ID of the target Voiceflow project.

name
string
required

Name of the evaluation.

Required string length: 1 - 100
enabled
boolean
required

When enabled, evaluations will be run on every new transcript.

prompt
string
required

This prompt describes how the LLM should apply this evaluation.

Required string length: 1 - 10000
type
enum<string>
required

The type of binary evaluations.

Available options:
boolean
truePrompt
string
required

This prompt describes which transcripts should evaluate as 'true'.

Required string length: 1 - 10000
falsePrompt
string
required

This prompt describes which transcripts should evaluate as 'false'.

Required string length: 1 - 10000
description
string | null

Description of the evaluation.

Maximum string length: 250
settings
object

The model settings passed when invoking this evaluation.

Response

201 - application/json
evaluation
object
required