StepAudio 2.5 ASR | EmpirioLabs AI Docs

StepFun · Transcription

POST /v1/audio/transcriptions

StepFun streaming speech recognition model for Chinese and English audio transcription.

At a glance

Field	Value
Model id	`stepaudio-2-5-asr`
Model release date	2026-04-24
Input modalities	Audio
Output modalities	Text
Context window	-
Weight precision	-
Region	International
Features	transcription, speech_to_text, streaming_asr, timestamps
Native inference	No
New	Yes
Supported endpoints	`POST /v1/audio/transcriptions`

Pricing

Charge	Spec	Rate
Transcription	per hour of audio	$0.022

Example request

$ curl https://api.empiriolabs.ai/v1/audio/transcriptions \
>   -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
>   -F model=stepaudio-2-5-asr \
>   -F file=@meeting.mp3

Parameters

Parameter	Type	Required	Default	Description
`file`	audio	no	-	Audio file upload for transcription.
`file_url`	string	no	-	Public URL to an audio file.
`audio_base64`	string	no	-	Base64 audio payload for JSON requests.
`language`	enum	no	`"en"`	Recognition language. · Allowed: `en`, `zh`
`hotwords`	array	no	-	Optional hotword list.
`enable_itn`	boolean	no	true	Enable inverse text normalization.
`enable_timestamp`	boolean	no	false	Return timestamp fields from StepFun SSE events when available.
`format`	enum	no	`"wav"`	Audio container format. · Allowed: `wav`, `mp3`, `ogg`, `pcm`
`codec`	string	no	-	PCM codec such as pcm_s16le.
`rate`	integer	no	`16000`	PCM sample rate in Hz. · Range: 8000 – 48000
`bits`	integer	no	`16`	PCM bit depth. · Range: 8 – 32
`channel`	integer	no	`1`	PCM channel count. · Range: 1 – 8

Notes

Supports wav, mp3, ogg, and pcm input. PCM requests should include codec, sample rate, bit depth, and channel count. Recognition language supports Chinese and English. Timestamp output is available through enable_timestamp.

Machine-readable schema: GET https://api.empiriolabs.ai/v1/models/stepaudio-2-5-asr.