Yandex SpeechKit technology overview

With the Yandex SpeechKit voice technologies, you can solve any task related to human speech. SpeechKit can recognize speech either in real time or from pre-recorded audio files while automatically detecting the speaker's language. It can also voice template phrases and long texts using the SpeechKit stock voices.

You can try the speech recognition and synthesis features under AI Speech in the AI Studio UI. Integrate these features into your application using the gRPC and REST APIs. For more on APIs in Yandex Cloud, see Yandex Cloud API concepts.

How to select a technology

Make your choice depending on your goal:

Goal

Getting started

Convert speech to text, including from an audio file or stream.

For an audio file, try recognition in the UI. For integration into an application or processing a voice stream, select recognition via the API.

Voice a ready-made text.

Try synthesis in the UI, explore the synthesis API, or synthesize speech via Realtime API if you need a streaming audio output.

Create an application that carries on a dialog, saves its context, and calls tools.

Explore voice agents and create an agent in the UI or connect it via Realtime API.

Manage speech processing and response logic yourself.

Combine recognition and synthesis in your application.

For requests to the SpeechKit recognition and synthesis API, the SpeechKit pricing policy applies. Synthesis via Realtime API is billed as per the Speech TTS Live model policy. You can find the cost of a voice agent, including audio processing, usage of the model and tools, in the AI Studio pricing policy.

Use cases

The table contains common SpeechKit use cases and their matching technologies.

Description Recommended technology Features and settings
Voice robot
Full or partial automation of phone conversations with customers. For a dialog with context and tools: Voice agent. For an application with a custom speech processing logic: Streaming recognition and synthesis using stock voices or brand voices.
Speech analytics
Quality control of agent performance
Transcribing and further analysis of audio recordings of conversations between customers and call center agents or robots. To recognize pre-recorded audio files: Asynchronous recognition of audio files.
Voice control in apps and smart devices
Voice assistant
The user requests an action or search using voice and the service responds with an action with a voice comment or an image. For a dialog with context and tools: Voice agent. For an application with a custom logic: Streaming recognition and synthesis using stock voices or brand voices.
Adjustments for visually impaired users
Voice control, voice guidance and comments for visually impaired users. For a user request: Streaming recognition.
For a system response: Speech synthesis using stock voices and brand voices.
Recognizing audio recordings made at meetings
Transcribing the audio recordings after the meeting is ended. To recognize pre-recorded audio files: Asynchronous recognition of audio files.
Voicing books and videos
Voicing a book or video with no human speaker involved. Speech synthesis using stock voices and brand voices.
Recording the minutes of a meeting
Transcribing the meeting minutes in real time. To recognize the participants' speech: Streaming recognition.
Video subtitles
Creating subtitles for recorded videos. To recognize an audio track: Asynchronous recognition of audio files.
Broadcast subtitles
Transcribing broadcasts in real time. To recognize the broadcast speech: Streaming recognition.
Transcribing voice messages
Converting short voice messages to text in messengers. To recognize audio files: Synchronous recognition. Recognition result settings.