AnalyticsConversations
Takes a human conversation into the module. Audio, audio with a ready transcript, or a transcript alone — all three are supported; audio itself is uploaded beforehand through `/api/files/upload-url` with file type `ConversationAudio`, so the file never travels through this API. A transcript without timings is accepted, and the response says so: talk share, silence, interruptions and speech rate stay unmeasured rather than being reported as zero.
Authorization
Bearer AuthorizationBearer <token>
JWT Authorization header using Bearer scheme
In: header
Request Body
application/json
TypeScript Definitions
Use the request body type in TypeScript.
Response Body
application/json
curl -X POST "https://test.api.speaknode.com/api/Analytics/conversations" \ -H "Content-Type: application/json" \ -d '{}'{
"session_id": "1ffd059c-17ea-40a8-8aef-70fd0307db82",
"duplicate": true,
"timings_missing": true,
"warning": "string"
}/Agents/{id}/public GET
Previous Page
Imports a zip of recordings and transcripts, uploaded beforehand as a file of type `ConversationArchive`. Recording and transcript are matched by name without extension (`call-001.mp3` with `call-001.json`); an optional `manifest.csv` or `manifest.json` in the root overrides the metadata; without it, telephony file names are read instead. Defaults to a dry run: the archive is validated and the report lists what would be created and what was rejected, without writing anything. Send it again with `dryRun: false` to import. POST
Next Page