Skip to main content


Tovie Platform provides a number of APIs for accessing its functions via HTTP in order to integrate them into third-party applications.

Dialog APIs

  • Chat API is an API for communicating with chatbots or customer service agents.
    Supported methods:

    • Sending client requests to the chat.
    • Fetching asynchronous events in the chat.
    • Fetching the chat message history.
    • Saving and loading the chat application state.
  • Operator API is an API for integrating Tovie Platform with any customer engagement platform.
    Supported methods:

    • Sending a message, a file, or any arbitrary data to a live agent.
    • Selecting the agent group where the message should be routed.
    • Subscribing to events in the communication channel.
  • Prompter API is an API for connecting prompters to Tovie Agent or integrating them into third-party services.
    Supported method:

    • Sending answers to users’ questions to an agent.


  • NLP Direct API is an API for accessing the NLU core, which is used for natural language understanding.
    Supported methods:

    • Text classification using a trained classifier.
    • Importing and exporting Tovie NLP data to and from existing Tovie Platform projects.
    • Adding, modifying, and deleting intents and entities.
    • Classifier training.
  • Model API is an API for connecting external NLU services to Tovie Platform.

Other APIs

  • Calls API is an API for working with outbound telephony.
    Supported methods:

    • Adding phone numbers with parameters to the outbound call campaign.
    • Canceling previously created call jobs.
    • Obtaining campaign reports.
    • Starting, pausing, resuming, and stopping the campaign.
  • Pushgate API is an API for managing bot events.
    Supported methods:

    • Creating pushbacks — special entities for triggering bot events.
    • Triggering bot events.
    • Scheduling events that will be triggered at a specified time.
  • Reporter API is an API for managing project analytics.
    Supported methods:

    • Receiving summary statistics on calls, unique users, and dialogs transferred to the agent.
    • Getting advanced statistics on the number of messages, sessions, and labels set.
    • Getting client lists, as well as blocking or unblocking them.
    • Getting a list of messages for specific sessions or clients.
    • Adding, deleting, and getting message and session labels.
    • Creating, activating, ending, and deleting experiments, as well as counting the number of sessions involved in experiments.
    • Generating reports on sessions, messages, clients, and calls, as well as deleting tasks for report generation.