Skip to main content
POST
/
inferenceResults
Create inference result
curl --request POST \
  --url https://api.galtea.ai/inferenceResults \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "sessionId": "session_123",
  "userId": "user_123",
  "index": 0,
  "status": "PENDING",
  "input": {},
  "actualOutput": "Model response text",
  "retrievalContext": "Retrieved context document",
  "latency": 123,
  "inputTokens": 123,
  "outputTokens": 123,
  "cacheReadInputTokens": 123,
  "tokens": 123,
  "costPerInputToken": 123,
  "costPerOutputToken": 123,
  "costPerCacheReadInputToken": 123,
  "cost": 123,
  "conversationSimulatorVersion": "<string>"
}
'
{
  "id": "ir_123",
  "sessionId": "session_123",
  "userId": "user_123",
  "index": 0,
  "status": "PENDING",
  "input": {
    "user_message": "User input text"
  },
  "actualOutput": "Model response",
  "retrievalContext": "<string>",
  "latency": 150,
  "inputTokens": 100,
  "outputTokens": 50,
  "cacheReadInputTokens": 20,
  "tokens": 150,
  "costPerInputToken": 0.00001,
  "costPerOutputToken": 0.00003,
  "costPerCacheReadInputToken": 0.000005,
  "cost": 0.001,
  "creditsUsed": 1,
  "conversationSimulatorVersion": "1.0.0",
  "traceId": "a1b2c3d4e5f6a1b2c3d4e5f6a1b2c3d4",
  "createdAt": "2023-11-07T05:31:56Z",
  "deletedAt": "2023-11-07T05:31:56Z"
}

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
​
sessionId
string

ID of the session

Example:

"session_123"

​
userId
string | null

ID of the user

Example:

"user_123"

​
index
integer

Index of the inference result

Example:

0

​
status
enum<string>

Status of the inference result

Available options:
PENDING,
GENERATED,
FAILED
​
input
object

Structured input data. For plain text, use { user_message: "..." }

​
actualOutput
string | null

Actual output text

Example:

"Model response text"

​
retrievalContext
string | null

Retrieval context information

Example:

"Retrieved context document"

​
latency
number | null

Latency in milliseconds

​
inputTokens
integer | null

Number of input tokens

​
outputTokens
integer | null

Number of output tokens

​
cacheReadInputTokens
integer | null

Number of cache read input tokens

​
tokens
integer | null

Total number of tokens

​
costPerInputToken
number | null

Cost per input token

​
costPerOutputToken
number | null

Cost per output token

​
costPerCacheReadInputToken
number | null

Cost per cache read input token

​
cost
number | null

Total cost

​
conversationSimulatorVersion
string | null

Version of the conversation simulator

Response

Inference result created successfully

​
id
string
Example:

"ir_123"

​
sessionId
string
Example:

"session_123"

​
userId
string | null
Example:

"user_123"

​
index
integer

Order index within the session

Example:

0

​
status
enum<string>
Available options:
PENDING,
GENERATED,
FAILED
Example:

"PENDING"

​
input
object

Structured input data. For plain text input, format is { user_message: "..." }

Example:
{ "user_message": "User input text" }
​
actualOutput
string | null
Example:

"Model response"

​
retrievalContext
string | null

The RAG retrieval context (retrieved documents/snippets) used to generate the actual output. Used by RAG-aware evaluation metrics.

​
latency
integer | null
Example:

150

​
inputTokens
integer | null
Example:

100

​
outputTokens
integer | null
Example:

50

​
cacheReadInputTokens
integer | null
Example:

20

​
tokens
integer | null

Total tokens

Example:

150

​
costPerInputToken
number | null
Example:

0.00001

​
costPerOutputToken
number | null
Example:

0.00003

​
costPerCacheReadInputToken
number | null
Example:

0.000005

​
cost
number | null
Example:

0.001

​
creditsUsed
integer | null
Example:

1

​
conversationSimulatorVersion
string | null
Example:

"1.0.0"

​
traceId
string | null

W3C trace ID for the root span created during direct inference. The same trace ID is propagated to the user endpoint via the traceparent header.

Example:

"a1b2c3d4e5f6a1b2c3d4e5f6a1b2c3d4"

​
createdAt
string<date-time>
​
deletedAt
string<date-time> | null