Skip to main content
POST
audio.transcribe-configured

Authorizations

Authorization
string
header
required

API key: sk_live_…

Headers

Idempotency-Key
string

Optional. A retry with the same key returns the stored result.

Maximum string length: 255
Pramaan-Max-Price-Micro
string

Optional. The most this request may be priced at, in micro-USD (a whole number above 0). It replaces the capability's own price ceiling for this request, lower or higher. Sources that could cost more are left out; when none is left the answer is 400 with the lowest limit that would work.

Pattern: ^[1-9]\d{0,14}$

Body

application/json
ref
string
required

The ref returned by audio.models.list.

Pattern: ^(ref_[0-9A-HJKMNP-TV-Z]{26}|rh_[A-Za-z0-9_-]{24,1024})$
audioUrl
string

Public HTTPS recording URL; source ceiling covers 10 minutes.

fileId
string

Id of an uploaded file holding the recording, up to 100 MB, in place of a public link.

Pattern: ^file_[0-9A-HJKMNP-TV-Z]{26}$
language
string

Spoken language tag; absent detects it.

Pattern: ^[a-z]{2,3}(?:-[a-z0-9]{2,8})*$
channels
integer
default:8

Maximum expected channel count, one to eight. Longer audio or extra channels are absorbed by the service.

Required range: 1 <= x <= 8
premium
boolean
default:false

Explicitly allow a premium or custom recognition model.

multilingual
boolean
default:false

Transcribe speech that switches between supported languages.

multichannel
boolean
default:false

Transcribe each channel independently.

speakerLabels
boolean
default:false

Label speakers within this recording.

speakerModel
enum<string>
default:latest

Speaker separation version. Default latest.

Available options:
latest,
first,
second
formatting
object

Transcript formatting controls.

keyTerms
string[]

Recognition hints for current models.

Maximum array length: 100

Term or phrase to recognize.

Required string length: 1 - 200
keywords
object[]

Weighted hints for older models.

Maximum array length: 100
replacements
object[]

Literal transcript replacements.

Maximum array length: 100
searchTerms
string[]

Terms to locate in the recording.

Maximum array length: 100

Phrase to locate.

Required string length: 1 - 200
redact
enum<string>[]

Sensitive entity types to redact.

Maximum array length: 50
Available options:
pci,
pii,
numbers,
credit_card,
credit_card_expiration,
credit_card_cvv,
ssn,
email_address,
phone_number,
name,
address,
date,
date_of_birth,
account_number,
bank_account,
driver_license,
passport,
ip_address,
url,
location,
organization,
medical_condition,
drug,
injury,
blood_type
utteranceGapSeconds
number
default:0.8

Silence separating utterances, in seconds.

Required range: 0.1 <= x <= 10
entities
boolean
default:false

Extract personal entities and contact details present in the speech.

summary
boolean
default:false

Summarize the submitted content.

topics
boolean
default:false

Identify topics in the content.

sentiment
boolean
default:false

Describe sentiment of text spans, without scoring a person.

intents
boolean
default:false

Identify expressed intentions in text spans.

customTopics
string[]

Up to 100 custom topic labels.

Maximum array length: 100

A topic to detect.

Required string length: 1 - 200
topicMode
enum<string>
default:extended

Strict returns only custom topics; extended also detects others.

Available options:
strict,
extended
customIntents
string[]

Up to 100 custom intent labels.

Maximum array length: 100

An intent to detect.

Required string length: 1 - 200
intentMode
enum<string>
default:extended

Strict returns only custom intents; extended also detects others.

Available options:
strict,
extended

Response

The result

transcript
object

Rich transcript, absent when stored as a result file.

resultFile
object

Full rich transcript as JSON when too large to return inline.