THE SPEECH & AUDIO API
Make audio
make sense.
Transcribe speech. Identify who said what.
Turn noisy recordings into something clear.
Good ideas start with
a conversation.
01 / THE MODEL CATALOG
Good models.
Great possibilities.
From a single recording to a live conversation.
Find the right model for what you’re building.
Whisper Large V3 Turbo
Fast multilingual speech recognition tuned for strong accuracy and lower latency.
Whisper Large V3
Multilingual speech recognition for accurate transcription, translation and timestamps.
Parakeet TDT 0.6B v3
Efficient multilingual ASR designed for high-throughput, low-latency transcription.
Whisper Realtime
Streaming multilingual speech recognition powered by the Whisper model family.
Nova-3
General-purpose speech recognition with language detection and detailed timestamps.
Nova-3 Realtime
Realtime speech recognition with fast partials and production-ready final transcripts.
Universal-3 Pro
High-accuracy speech recognition for multilingual audio and long-form transcription.
Universal-3 Realtime Pro
Low-latency streaming ASR for live conversations and responsive voice applications.
MOSS Transcribe 1.0
General speech transcription optimized for clear, structured text output.
MOSS Transcribe Diarize Pro
Speech transcription with speaker-aware segments for conversations and meetings.
Fun-ASR Flash
Fast multilingual speech recognition for short audio and interactive workloads.
Qwen Audio 3.0 ASR Flash
Multilingual audio understanding and fast speech-to-text transcription.
Qwen Audio 3.0 ASR Flash Streaming
Realtime multilingual ASR with incremental transcripts for streaming audio.
Qwen3 Forced Alignment
Aligns an existing transcript to speech and returns precise word-level timestamps.
DeepFilterNet3
Neural speech enhancement that suppresses noise while preserving voice clarity.
Silero VAD
Detects speech regions in audio for trimming, segmentation and preprocessing.
Explore all deployments, languages, parameters and model-specific prices. Open Console when you’re ready to run a model.
02 / BUILT FOR DEVELOPERS
Your next feature.
A few lines away.
Pick a model, try it in the playground, then bring the same endpoint into your application.
curl --request POST \
"https://api.speechinfra.com/v1/inference/whisper-large-v3-turbo-1" \
--header "Authorization: Bearer $SPEECH_API_KEY" \
--form "file=@conversation.wav"{
"text": "Good ideas start with a conversation."
}SPEECHINFRA, EXPLAINED
Speech models. One integration.
Speechinfra is an API platform for speech-to-text, realtime speech recognition, speaker diarization, speech alignment, audio enhancement and voice activity detection. Choose a model and use its endpoint with one project API key. Each model has its own supported languages, settings and price.
Transcripts & subtitles
Turn short recordings into text. Choose a model with timestamps when you need to align words or build subtitles.
Conversations & speakers
Use speaker-aware models for dialogue clips or realtime ASR for streaming speech. Speaker and language support vary by model.
Audio preparation
Reduce background noise, detect speech regions or align an existing transcript with its recording.
Start with a short clip.
Create and verify your account, copy your project key, then send a short file to your chosen endpoint. Our direct file API currently supports up to 30 seconds and 20 MiB per request. A funded balance is required.
Read the API quickstart →A few things worth knowing.
What is Speechinfra?
Speechinfra provides API access to speech and audio models for transcription, realtime speech recognition, speaker diarization, speech alignment, enhancement and voice activity detection. Choose a specific model for each request.
How do I start?
Create an account and verify your email. Copy your project API key from Console for server-side API calls. The web playground uses your signed-in session; you do not paste a key into it. A funded balance is required to run a model.
Which files can I send?
The direct file API accepts short audio or video clips, up to 30 seconds and 20 MiB. WAV is used in our examples. Accepted containers and additional settings depend on the selected model. Realtime uses binary mono PCM16 at 16 kHz, not uploaded compressed files.
Which languages and speaker features are supported?
Support varies by model. Each public model page lists the language codes, automatic language detection support and request parameters exposed by our API. Do not assume every model supports every language, speaker labels or word timestamps.
How is usage billed?
Prices are in USD per minute of audio and vary by model. For file processing, the server measures duration and reserves the cost before starting the model. Insufficient funds return HTTP 402. Failed file processing releases the reservation. Realtime reserves incrementally and charges for audio already sent.
Can I use the API without opening Console each time?
Yes. Send your project key in the Authorization: Bearer header from your application backend. Never expose provider credentials or embed a project key in a public frontend bundle.
LESS INFRASTRUCTURE. MORE IDEAS.