99 languages
Automatic language detection included.
AI Speech Technology
Transcribe audio with clarity and precision.
The ACKustics Speech-to-Text API delivers accurate transcriptions in 99 languages, with word-level timestamps and scalable processing for production-grade applications.
Word-level timestamps
Every word, exactly when it was spoken.
OpenAI-compatible API
Switch with a one-line change.
Built for scale
Handle more audio with confidence.
STT now.
TTS next.
We're building a speech platform for developers. Start with Speech-to-Text today and add Text-to-Speech when you're ready.
Speech-to-Text
Available now- Accurate transcription
- Word-level timestamps
- 99 languages with automatic detection
- Plain text and JSON output
Text-to-Speech
Coming soon- Natural, expressive voices
- High-quality synthesis
- Multiple voice options
- Flexible & customizable
From audio to text in three steps
Send your audio
Upload an audio file through the API and choose the options you need.
We transcribe it
The language is detected automatically, and every word gets its own timestamps.
Get your transcript
Receive plain text or structured JSON, ready for search, captions or downstream automation.
99 languages
Transcribe speech in 99 languages, from the most widely spoken to regional ones. Language detection is automatic, or you can set it yourself.
Accuracy varies by language and audio quality.
A drop-in API for your stack.
Already using an OpenAI-style transcription endpoint? Point your client at ACKustics, change the key and keep your code. Request word-level timestamps with a single parameter.
- OpenAI-compatible transcription endpoint
- Word-level timestamps on request
- Automatic or explicit language selection
# Transcribe a file with word-level timestamps
curl https://api.ackustics.com/v1/oai/audio/transcriptions \
-H "Authorization: Bearer $ACKUSTICS_API_KEY" \
-F file=@meeting.mp3 \
-F model=whisper3-large-v3 \
-F language=auto \
-F response_format=json \
-F "timestamp_granularities[]=word"Built for production workloads
High accuracy
Modern speech models deliver precise transcriptions, even in challenging audio.
Pay for what you use
Prepaid credits spent by usage, with the rate shown before each operation. No subscriptions and no minimum commitment.
Clear data retention
Retention depends on the feature and access channel, and is disclosed before you run an operation.
Built for scale
Scalable infrastructure and reliable APIs designed for high-volume workloads.
Built for how you work
Media & content
Transcribe interviews, podcasts and video for editing and publishing.
Captions & subtitles
Use word-level timestamps to build accurate, synchronized captions.
Audio search
Turn archives into searchable text and jump straight to the right moment.
Voice AI
Power assistants, voice apps and smart experiences.
Frequently asked questions
Which languages are supported?
99 languages. The language is detected automatically, or you can set it explicitly in the request. Accuracy varies by language and audio quality.
How does billing work?
You buy prepaid credits and spend them as you use each service. The rate depends on the service and model, and is shown before each operation. Speech-to-Text is billed per second of audio processed. There are no subscriptions. See the pricing page for details.
Do you store my audio?
Retention depends on the feature and the access channel you use, and is disclosed before you run an operation. Our Privacy policy describes the details.
What output formats are available?
Plain text and JSON. Word-level timestamps are available on request.
Can I get volume pricing or a custom model?
Yes. We offer special pricing for high volumes and can discuss custom model deployments. Contact sales.
How do I get access?
ACKustics is in pre-launch. Request a demo and we will get in touch with the next steps.
Ready to unlock the value of your audio?
Build better experiences with accurate, timestamped transcriptions that scale with you. Need volume pricing or a custom model? Talk to us.