Skip to main content
POST
/
api
/
sre
/
llm
/
v1
/
evaluate
Evaluate
curl --request POST \
  --url https://developer.synq.io/api/sre/llm/v1/evaluate \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "outputSchema": {},
  "systemPrompt": "<string>",
  "messages": [
    {
      "content": "<string>",
      "role": "MESSAGE_ROLE_UNSPECIFIED"
    }
  ],
  "modelType": "MODEL_TYPE_UNSPECIFIED"
}
'
{
  "output": {},
  "metrics": {
    "inputTokens": 123,
    "outputTokens": 123,
    "totalTokens": 123,
    "latencyMs": 123,
    "model": "<string>",
    "cacheWriteTokens": 123,
    "cacheReadTokens": 123
  }
}

Authorizations

​
Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

Request to evaluate an LLM request with structured output.

​
outputSchema
object
required

JSON schema defining the structure of the expected output. The LLM will produce output conforming to this schema.

​
systemPrompt
string
required

Main system prompt providing instructions to the LLM. This should be constant as it will be cached for efficiency.

​
messages
Message · object[]

History of messages in the conversation. Must contain at least one message. The last message is used as the final request to the LLM.

Minimum array length: 1
​
modelType
enum<string>

Type of model to use for the evaluation. Defaults to MODEL_TYPE_SUMMARY if not specified or set to MODEL_TYPE_UNSPECIFIED.

Available options:
MODEL_TYPE_UNSPECIFIED,
MODEL_TYPE_SUMMARY,
MODEL_TYPE_THINKING

Response

200 - application/json

Success

Response from the LLM evaluation.

​
output
object

Structured output from the LLM conforming to the provided schema.

​
metrics
LlmResponseMetrics · object

Metrics about the LLM response.