Skip to main content
POST
cURL

Authorizations

Authorization
string
header
required

Your API key: Authorization: Bearer aircaps_sk_...

Headers

Content-Type
string

The clip's media type, for example audio/wav. Informational: the format is detected from the bytes.

Query Parameters

speaker_ids
string

Only these speakers: comma-separated, or the parameter repeated. Default: every enrolled speaker of the account. Verification of one person is speaker_ids=<their id>.

Example:

"spk_01k73a9x8w7v6u5t4s3r2q1p0n,spk_01k73a7m6n5b4v3c2x1z0l9k8j"

top_k
integer

Return only the best top_k results (all speakers are still scored; margin and speakers_compared are unchanged).

Required range: 1 <= x <= 1000
model
enum<string>

Optional: aircaps-speaker-id-v1. Any other value returns 400 unsupported_model.

Available options:
aircaps-speaker-id-v1

Body

The clip's raw audio bytes (any common audio or video format; channels are mixed down and audio is resampled to 16 kHz). Not multipart form data. The whole clip is used as given: crop it to one person's speech first.

The body is of type file.

Response

The scores.

id
string
required

Identification ID (identifies the request; identifications are not stored).

Example:

"sid_01k73c0q9w8e7r6t5y4u3i2o1p"

object
enum<string>
required
Available options:
speaker_identification
model
enum<string>
required
Available options:
aircaps-speaker-id-v1
created_at
string<date-time>
required
audio_duration
number
required

Seconds of audio in the clip.

Example:

4.21

speakers_compared
integer
required

Speakers the clip was scored against (before top_k).

Example:

3

margin
number | null
required

Best score minus the second best. null with one speaker.

Example:

0.476

results
object[]
required

Best first (ties by speaker ID).