Yandex Cloud
Search
Contact UsGet started
  • Blog
  • Pricing
  • Documentation
  • All Services
  • System Status
    • Featured
    • Infrastructure & Network
    • Data Platform
    • Containers
    • Developer tools
    • Serverless
    • Security
    • Monitoring & Resources
    • ML & AI
    • Business tools
  • All Solutions
    • By industry
    • By use case
    • Economics and Pricing
    • Security
    • Technical Support
    • Customer Stories
    • Start testing with double trial credits
    • Cloud credits to scale your IT product
    • Gateway to Russia
    • Cloud for Startups
    • Education and Science
    • Yandex Cloud Partner program
  • Blog
  • Pricing
  • Documentation
© 2025 Direct Cursus Technology L.L.C.
Yandex SpeechSense
  • Getting started
    • All guides
      • Creating a connection
  • Audit Trails events
  • Access management
  • Pricing policy
  • Release notes
  • FAQ

In this article:

  • Creating a connection for an audio
  • Creating a connection for a chat
  1. Step-by-step guides
  2. Connections
  3. Creating a connection

Creating a connection

Written by
Yandex Cloud
Updated at May 14, 2025
  • Creating a connection for an audio
  • Creating a connection for a chat

You can upload a two-channel audio or text chat to SpeechSense connections. To create a connection, you need either the Admin or Editor role for the space.

Creating a connection for an audioCreating a connection for an audio

  1. Open the SpeechSense home page.

  2. Go to the space of your choice.

  3. In the top-right corner, click More →  Create connection.

  4. Name the new connection and, optionally, enter a description.

  5. Select the Two-channel audio data type.

  6. Under Agent, Customer:

    1. Specify the channels with the agent's voice and customer's voice. The left channel corresponds to audio track 0, while the right channel corresponds to audio track 1.

    2. Specify agent and customer keys from the metadata file. This file contains the call information collected from CRM systems, PBXs, or other sources.

      By default, connections include keys with the name and ID of the agent and the customer. In the Name in the system field, enter the name the key will have in SpeechSense. You can also add a description, if required.

      To specify additional metadata for the agent and the customer, click Add key. Make sure to enter the new key name and select a metadata type.

  7. Under Shared metadata, provide the keys from the metadata file that are not related to the agent or the customer.

    By default, connections include required keys with the date, direction, and language of the call. In the Name in the system field, enter the name the key will have in SpeechSense. You can also add a description, if required.

    To specify additional metadata, click Add key. Make sure to enter the new key name and select a metadata type.

    SpeechSense can analyze audio recordings in Russian and Kazakh. First it recognizes your recording in Kazakh and then translates the transcript into Russian for further analysis. Each audio metadata must contain a string indicating the languages supposedly used in the conversation. When uploading an audio file in Kazakh, provide the language: ru-RU, kk-KZ string in the metadata. For Russian, provide language: ru-RU. Language codes are case-insensitive.

    To upload related dialogs into the connection, add the ticket_id string key. The dialogs will be linked by this key. Enter the name the key will have in SpeechSense, e.g., Task number. You can also add a description, if required.

  8. Click Create connection.

Creating a connection for a chatCreating a connection for a chat

  1. Open the SpeechSense home page.

  2. Go to the space of your choice.

  3. In the top-right corner, click More →  Create connection.

  4. Name the new connection and, optionally, enter a description.

  5. Select the Сhat data type.

  6. Under Agent, Customer, Bot, specify the keys from the metadata file. This file contains the dialog information collected from chats, CRM systems, or other sources.

    By default, connections include keys with agent's, customer's, and bot's names and IDs. In the Name in the system field, enter the name the key will have in SpeechSense. You can also add a description, if required.

    To specify additional agent, customer, and bot metadata, click Add key. Make sure to enter the new key name and select a metadata type.

  7. Under Shared metadata, provide the keys from the metadata file that are not related to the agent, customer, and bot.

    By default, keys with date, direction, and dialog language are added to the connection. In the Name in the system field, enter the name the key will have in SpeechSense. You can also add a description, if required.

    To specify additional metadata, click Add key. Make sure to enter the new key name and select a metadata type.

    SpeechSense can analyze dialogs in Russian and Kazakh; it translates the transcript into Russian for further analysis. Each dialog metadata must contain a string indicating the languages supposedly used in the conversation. When uploading dialogs in Kazakh, provide the language: ru-RU, kk-KZ string in the metadata. For Russian, provide language: ru-RU. Language codes are case-insensitive.

    To upload related dialogs into the connection, add the ticket_id string key. The dialogs will be linked by this key. Enter the name the key will have in SpeechSense, e.g., Task number. You can also add a description, if required.

  8. Click Create connection.

Was the article helpful?

Previous
Removing a user from a space
Next
Creating a project
© 2025 Direct Cursus Technology L.L.C.