Skip to main content
POST
Upload a recording

Authorizations

Authorization
string
header
required

Your workspace secret key, sk_live_ followed by 32 hex characters. Create one in the console under Developers.

Headers

Idempotency-Key
string

Up to 255 characters. The same key within 24 hours replays the first answer and creates nothing new. Build it from your own record, such as order-8842-reminder-1.

Maximum string length: 255

Body

multipart/form-data
file
file
required

The recording. One file per request.

agent_id
string

The agent whose analysis settings to use. Leave out for a transcript only.

Example:

"agt_01JD7Q2M4N6P8R0S2T4V6W8X0Y"

language
string

auto (default), hi, gu, en, mr, pa, bn, ta, te, kn, ml or ur.

recorded_at
string

When the conversation happened, so "tomorrow" becomes the right date.

webhook_url
string

Where to send the result event. https only.

metadata
string

Your own ids, as a JSON object of strings. Up to 16 keys.

Response

Accepted

An uploaded recording. Lists carry it without transcript, analysis and delivery; GET /v2/recordings/{id} adds them.

agent_id
string | null
required
agent_revision
integer | null
required
bytes
integer
required
Required range: x >= 0
channels
enum<integer> | null
required
Available options:
1,
2,
null
completed_at
string<date-time> | null
required

When it completed or failed.

cost
object
required

Transcription ₹0.30 and insight ₹0.20 per minute of audio, billed per second rounded up to a paisa, each charged once when its stage succeeds.

created_at
string<date-time>
required
detected_language
string | null
required

The main language spoken, once transcribed.

duration_secs
number | null
required

The recording's length, measured; null until known.

Required range: x >= 0
error
object | null
required
file_name
string
required
id
string
required

The recording id; also data.call.id of its webhook and the reference of its wallet transactions (kind recording).

Pattern: ^rec_[0-9A-Z]{26}$
insight
boolean
required

true when an agent's summary and analysis were asked for (agent_id).

language
string | null
required

The language you asked for; null for automatic.

object
any
required
progress
object | null
required

While transcribing: parts transcribed of the total (a long two-channel recording is transcribed in parts).

queue_position
integer | null
required

While queued: about how many recordings are ahead of it.

Required range: x >= 0
recorded_at
string<date-time> | null
required
source
enum<string>
required
Available options:
upload,
url
speakers
enum<string> | null
required

How the transcript's speakers were told apart: by channel, automatically from one channel, or not at all (every line Speaker). null until transcribed.

Available options:
channels,
diarized,
unlabelled,
null
started_at
string<date-time> | null
required

When it left the queue.

status
enum<string>
required

queued (waiting for a slot; uploads run with their own capacity so calls are never slowed) -> transcribing -> analyzing (with an agent) -> completed | failed.

Available options:
queued,
transcribing,
analyzing,
completed,
failed
transcribed_at
string<date-time> | null
required
analysis
object | null

GET by id only: the agent's analysis, as a call's; null until it exists.

delivery
object | null

GET by id only: the result's delivery to webhook_url.

metadata
object
transcript
object[]

GET by id only, once transcribed.