Voice Transcription API

Long Form ASR based on deep full-sequence convolutional neural networks, converting long audio data (within 5 hours) into text data.

byiFlytek Open Platform

200+calls

N/Alatency

Updated: Oct 27, 2026, 12:40 AM

Continue setup in your AI

Only public catalog IDs and a safe return path are handed to the Plugins page. This does not run the tool or confirm current availability, authorization, price, limits, or upstream status; those conditions are checked again before execution.

Selected tool: xfyun_api.voice_transcription.v2
Provider scope: iFlytek Open Platform
Catalog updated: 2026-10-27T00:40:01.989419+00:00
Business task: Pending verification

Continue setupYou can return to this tool page if the selection is incomplete or expires.

Run on this page: Execute is the in-site test path. Continue setup is for adding the public selection to an external AI; the two paths are tracked and validated separately.

Discover -> Inspect -> Call

Call setup

Tool IDxfyun_api.voice_transcription.v2

Request parametersJSON

Before this Call

1 credit / call

The final charge is settled from actual provider usage and will appear in Usage.

Field reference

pdstringOptional

Domain personalization parameter

hotWordstringOptional

Hot words to improve recognition rate of professional vocabulary

roleNumintegerOptional

Number of speakers, range 0-10, default 0 for blind separation

audioUrlstringOptional

Audio URL external link address

durationintegerRequired

Audio real duration

fileNamestringRequired

Audio file name, preferably with real file extension

fileSizeintegerRequired

Audio file size (bytes)

languagestringOptional

Language type

roleTypeintegerOptional

Whether to enable role separation: 0 - off (default), 1 - on

audioModestringOptional

Audio upload method: fileStream (default), urlLink

candidateintegerOptional

Multi-candidate switch: 0 - off (default), 1 - on

trackModeintegerOptional

Track-by-track transcription mode: 1 - no track separation (default), 2 - dual track separation mode

transModeintegerOptional

Translation mode: 2 - by paragraph (currently only supports paragraph translation)

callbackUrlstringOptional

Callback URL to notify when task is completed

standardWavintegerOptional

Whether it is standard pcm/wav (16k/16bit/mono): 0 - non-standard (default), 1 - standard

languageTypeintegerOptional

Language recognition mode selection (for cn language): 1 - automatic Chinese-English mode (default), 2 - Chinese mode, 4 - pure Chinese mode

transLanguagestringOptional

Language to translate to (must be different from transcription language)

Continue setup in your AI

Selected tool

xfyun_api.voice_transcription.v2

Provider scope

iFlytek Open Platform

Catalog updated

2026-10-27T00:40:01.989419+00:00

Business task

Pending verification

Continue setupYou can return to this tool page if the selection is incomplete or expires.