
Yandex SpeechKit
Speech technologies based on machine learning to create voice assistants, automate call centers, monitor service quality, and perform other tasks.
Leverage the advanced technology behind the wildly successful Alice voice assistant, now ready for use in your business.
Instant speech recognition
In a fraction of a second, SpeechKit accurately recognizes speech, allowing our clients' voice assistants to communicate quickly and easily.
Realistic voices
Every phrase generated by Yandex SpeechKit sounds natural thanks to human-like pronunciation and real-time speech synthesis.
SpeechKit Brand VoiceNew
Choose the right version for you: the full version creates a smart voice assistant while the adaptive version gives your brand a unique voice in just a month.
Analytics
Yandex SpeechKit helps improve customer service by analyzing 100% of calls. Get your ear to the ground to keep your business processes ahead of the competition.
Transparent pricing
You have full control over your expenses, with costs for audio recognition calculated based on the length of the track, and speech synthesis by character count.
SpeechKit HybridNew
A solution for the most demanding customers who need to control speech processing and synthesis within their own infrastructure.
SpeechKit Hybrid
SpeechKit’s ML models can now be deployed to your infrastructure. We offer both hybrid options and 100% on-premise deployments for sensitive traffic.
A unique voice for your brand with SpeechKit Brand Voice
A distinct voice that fits your brand’s communication strategy. We can create a voice synthesis based on your speaker with exclusive access to the voice model for use with:
Voice assistants
Sophisticated dialogs
Non‑standard cases
FAQ
How do I use Yandex SpeechKit?
The service runs via the HTTP API. You can find instructions for using the service in the documentation and either start using the service yourself or contact us. We’ll help you find a partner that will develop a solution specifically for your project.
Why should I register in the Yandex Cloud console?
To use the API, you need to get an ID (an IAM token or API key). This ID is linked to your account in Yandex Cloud.
What is a recognition model?
A recognition model is a neural network that is trained to recognize speech in a specific language. The models are trained on datasets generated by Yandex services and applications, allowing us to continually improve the quality of speech recognition.
What audio formats can Yandex SpeechKit recognize?
The service can recognize audio in MP3, LPCM and OggOpus formats.