Skip to main content
POST
Revise metric configuration with AI (Update With AI)
You are reading the docs of Galtea 5.12. Read this page in the latest version.

Authorizations

Authorization
string
header
required

API key authorization. Pass your API key in the Authorization header as a Bearer token. Both new (gsk_*) and legacy (gsk-) API keys are accepted, e.g. Authorization: Bearer gsk_... or Authorization: Bearer gsk-....

Body

application/json
name
string
required

The metric's user-authored name — context for the revision, not returned in the response.

Required string length: 1 - 200
Example:

"Refusal Quality"

evaluationType
enum<string>
required

PARTIAL_PROMPT (binary LLM judge) or HUMAN_EVALUATION (0-100 annotator guidelines).

Available options:
PARTIAL_PROMPT,
HUMAN_EVALUATION
currentJudgePrompt
string
required

The metric's current judge prompt — the starting point the revision builds on.

Required string length: 1 - 20000
feedback
string
required

Free-text instruction describing what to change about the current judge prompt.

Required string length: 1 - 2000
Example:

"it should not penalise a short answer"

description
string

The metric's user-authored description — the criteria the judge prompt must evaluate.

Required string length: 1 - 5000
currentEvaluationParams
string[]

The metric's current evaluation parameters, changed only if the feedback requires it. Always send them when the metric has any; omitting them can produce a revision whose declared parameters do not match its prompt.

Maximum array length: 20
Required string length: 1 - 64
tags
string[]

The metric's user-provided tags — context for the revision, not returned in the response.

Maximum array length: 20
Required string length: 1 - 64
productId
string

Optional product ID for additional context

Minimum string length: 1
Example:

"prod_123"

specificationIds
string[]

At most one specification id, forwarded as optional context.

Minimum string length: 1

Response

Revised metric configuration suggestion

judgePrompt
string
required

The revised judge prompt, built on the current one.

evaluationParams
string[]
required

The revised evaluation parameters.

description
string | null
required

The revised metric description.