Sessions Overview

A session ties audio capture, processing and results together under one session_id. Base path: /voice/v1 on https://api.eka.care. All endpoints except Discovery need Authorization: Bearer YOUR_ACCESS_TOKEN (get a token).

1. Create Session   POST /voice/v1/sessions                        → session_id, upload_url
2. Send audio       POST /voice/v1/sessions/{id}/audio/{file}     (single/chunked, HTTP)
                    or wss://… WebSocket                           (stream)
3. End Session      POST /voice/v1/sessions/{id}/end               → 202, processing starts
4. Get Session      GET  /voice/v1/sessions/{id}                  (poll ~1s until ≠202)
                    → transcript + structured notes per template

API	Endpoint
Create Session	`POST /voice/v1/sessions`
Upload Audio	`POST /voice/v1/sessions/{session_id}/audio/{file_sequence}`
Stream Audio	`wss://…` (from `upload_url`)
End Session	`POST /voice/v1/sessions/{session_id}/end`
Get Session	`GET /voice/v1/sessions/{session_id}`
Update Session	`PATCH /voice/v1/sessions/{session_id}`
Upload Transcript	`POST /voice/api/v1/transaction/{session_id}/convert-to-template`
Process Template	`POST /voice/v1/sessions/{session_id}/process/template/{template_id}`

Already have the transcript? Upload Transcript is an optional alternative to steps 2–3 — post the transcript and a template_id straight to the session, then poll Get Session for the structured note. Also see Templates (define note formats) and Documents (read and edit generated notes).

SDKs

API Reference

Resources