Alibaba Cloud: Fun-ASR

Fun-ASR converts recorded speech into text for searchable audio archives. It is useful for turning interviews, voice notes, and recorded conversations into material that can be reviewed, summarized, or indexed.

Compare models

fun-asr

AvailableAlibaba CloudSpeech to TextSync / Async

Price: From $0.000035
Modalities: Audio

About Fun-ASR

Fun-ASR is Alibaba's speech recognition model for recorded audio, called asynchronously on files. Within the Fun-ASR line it is the general file-transcription choice, as opposed to the Realtime streaming model and the short-clip Flash release. Alibaba lists context enhancement, custom hotwords, speaker diarization, sensitive-word filtering, and sentence- or word-level timestamps for it, which makes it suited to meeting and interview archives.

Where it works well

Speaker diarization labels who said what in multi-person recordings such as meetings and interviews.
Custom hotwords bias recognition toward product names, people, and jargon that a general vocabulary would miss.
Sentence- and word-level timestamps allow transcripts to be matched back to the audio for subtitles or search.
Accepts mainstream audio formats at any sample rate, so recordings rarely need resampling first.

When to choose another model

It processes finished files asynchronously, so it cannot drive live captions; Fun-ASR Realtime fits that job.
Output is transcript text only; summaries, translation, or action items need a separate language model.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Speech to textTokenLab endpoint

POST/v1/audio/transcriptions

curl -X POST "https://api.tokenlab.sh/v1/audio/transcriptions" \
  -H "Authorization: Bearer sk-xxx" \
  -F file="@audio.mp3" \
  -F operation="stt" \
  -F response_format="json" \
  -F model="fun-asr" \
  -F language="en"

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Speech to text

per second

Official price: $0.000035
Verified price: —
Discount: —

	Official priceper second	Verified priceper second	Discount
Speech to text	$0.000035	—	—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Oct 14, 2026, 11:11 PM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Fun-ASR in Console with a prompt ready to edit or send.

Help me try fun-asr with a short audio request at /v1/audio/transcriptions. Show the result and cost.

Use cases

Meeting minutes
Transcribe a recorded call with speaker labels, then pass the text to a language model to draft minutes and action items.
Interview archives
Turn hours of recorded interviews into searchable text with word timestamps, so a reviewer can jump to a quoted passage.
Brand-term transcripts
Add product and company names as hotwords so recorded webinars come out with the right spelling.

Prompt examples

Transcribe this 90-minute board meeting recording, label each speaker, and keep sentence timestamps.

Transcribe this customer interview with the hotwords 'Kubernetes', 'roadmap', and 'Q3 churn'.

Transcribe this podcast episode and filter sensitive words from the text output.

FAQ

Is Fun-ASR for live or recorded audio?

Recorded audio. It is called asynchronously on files. For microphone input or live captions use Fun-ASR Realtime, which consumes audio as a session progresses and returns text while people keep talking.

Can Fun-ASR tell speakers apart?

Yes. Alibaba lists speaker diarization among its features, so the transcript can separate speakers in a meeting or interview. Accuracy depends on audio quality and how much speakers overlap.

Does Fun-ASR return timestamps?

Yes. Alibaba documents sentence-level and word-level timestamps for it, which you can use to build subtitles or to link transcript text back to a position in the recording.

How large a file can Fun-ASR handle?

Alibaba's file-recognition documentation allows recordings up to 2 GB and 12 hours. Longer material should be split. Mainstream formats such as AAC, WAV, and MP3 are accepted.

How much does Fun-ASR cost?

On TokenLab, Fun-ASR costs $0.000035 per second. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

Which endpoint should Fun-ASR use?

Use https://api.tokenlab.sh/v1/audio/transcriptions for Fun-ASR. The request example below shows the matching code shape.

Which operations does Fun-ASR support?

Fun-ASR supports Speech to text. Select an operation above to see its endpoint and request example.