logoSpeaknode
Sign Up
AnalyticsConversationsSessionid

Corrects what the pipeline could not know: the performer, the language, the customer, and which speaker is the operator. Re-assigning speakers recomputes the time-based metrics straight away — that is arithmetic, no model is asked, and a talk share attributed to the wrong side should not survive until the next full re-analysis.

PUT
/Analytics/conversations/{sessionId}

Authorization

Bearer
AuthorizationBearer <token>

JWT Authorization header using Bearer scheme

In: header

Path Parameters

sessionId*string

Conversation to correct.

Formatuuid

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

Response Body

application/json

curl -X PUT "https://test.api.speaknode.com/api/Analytics/conversations/497f6eca-6276-4993-bfeb-53cbbbba6f08" \  -H "Content-Type: application/json" \  -d '{}'
{
  "success": true,
  "error": "string",
  "records_updated": 0
}

The conversation card: transcript with timecodes, everything measured, and a signed link to the audio. A metric that could not be measured comes back marked as such rather than as zero — a zero talk share reads as "the rep never spoke". GET

Previous Page

Puts a conversation through the pipeline: recognition when it has only audio, measurement once a transcript exists. Works for an uploaded recording and for our own agent session alike — the latter gets its transcript laid out from the chat history first. POST

Next Page