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 Foundation Models
    • All tutorials
    • Disabling request logging
    • Getting an API key
    • Batch processing
      • Creating a simple assistant
      • Creating a search assistant
      • Creating a search assistant specifying file metadata and indexes
      • Getting intermediate response generation results
  • Yandex Cloud ML SDK
  • Compatibility with OpenAI
  • Access management
  • Pricing policy
  • Public materials
  • Release notes

In this article:

  • Getting started
  • Create an assistant
  1. Step-by-step guides
  2. AI Assistant API
  3. Getting intermediate response generation results

Creating an assistant with intermediate response generation results

Written by
Yandex Cloud
Updated at April 11, 2025
  • Getting started
  • Create an assistant

The AI Assistant API feature is at the Preview stage.

AI Assistant API is a tool for creating AI assistants. It can be used to create personalized assistants, implement a generative response scenario adapted based on external information (known as retrieval augmented generation, or RAG), and save the model's request context. When making requests to the assistant, you can get intermediate generation results as the model is generating a response.

Getting started

To use the examples:

SDK
  1. Create a service account and assign the ai.assistants.editor and ai.languageModels.user roles to it.

  2. Get the service account API key and save it.

    The following examples use API key authentication. Yandex Cloud ML SDK also supports IAM token and OAuth token authentication. For more information, see Authentication in Yandex Cloud ML SDK.

  3. Use the pip package manager to install the ML SDK library:

    pip install yandex-cloud-ml-sdk
    

Create an assistant

This example shows how to create an assistant to store your conversations with YandexGPT Pro RC. When you access the model, your assistant will output intermediate generation results as the model is generating a response.

SDK
  1. Create a file named simple-assistant.py and paste the following code into it:

    #!/usr/bin/env python3
    
    from __future__ import annotations
    
    from yandex_cloud_ml_sdk import YCloudML
    
    
    def main() -> None:
        sdk = YCloudML(
            folder_id="<folder_ID>",
            auth="<API_key>",
        )
        sdk.setup_default_logging()
    
        assistant = sdk.assistants.create(
            "yandexgpt",
            temperature=0.5,
            max_prompt_tokens=50,
            ttl_days=6,
            expiration_policy="static",
        )
    
        thread = sdk.threads.create(
            name="foo",
            ttl_days=6,
            expiration_policy="static",
        )
    
        # Basic cycle of user interaction with the assistant
        input_text = input('Ask a question to the assistant (or "exit" to exit): ')
        while input_text.lower() != "exit":
    
            thread.write(input_text)
    
            # This way you can give the whole thread contents to the model
            run = assistant.run_stream(thread)
    
            # This way you can see the intermediate results as the model generates a response
            for event in run:
                print(event._message.parts)
    
            # This way you can see all fields of the final result
            print(f"run {event=}")
    
            input_text = input('Ask the assistant your question (or "exit" to exit): ')
    
        # Displaying the entire history of messages in the chat room
        print("Outputting the whole message history when exiting the chat:")
        for message in thread:
            print(f"    {message=}")
            print(f"    {message.text=}\n")
    
        # Deleting everything you do not need anymore
        for assistant in sdk.assistants.list():
            assistant.delete()
    
        thread.delete()
    
    
    if __name__ == "__main__":
        main()
    

    Where:

    • <folder_ID>: ID of the folder in which the service account was created.

    • <API_key>: Service account API key you got earlier required for authentication in the API.

      The following examples use API key authentication. Yandex Cloud ML SDK also supports IAM token and OAuth token authentication. For more information, see Authentication in Yandex Cloud ML SDK.

  2. Run the created file:

    python3 simple-assistant.py
    

    The example implements the simplest chat possible: enter your requests to the assistant from your keyboard and get answers. To end the dialog, enter exit.

    Approximate result
    Ask the assistant a question (or "exit" to exit): Hi!
    ('Hello',)
    ('Hello! What can I do for you?',)
    ('Hello! What can I do for you?',)
    run event=RunStreamEvent(status=<StreamEvent.DONE: 3>, error=None, _message=Message
    (id='fvtbkt2tbf7a********', parts=('Hello! What can I do for you?',), thread_id='fvt50ma5302n********', 
    created_by='ajegtlf2q28a********', created_at=datetime.datetime(2025, 3, 13, 17, 51, 22, 146833), 
    labels=None, author=Author(id='fvtle31p6lv2********', role='ASSISTANT'), citations=()))
    Ask your question to the assistant (or "exit" to exit): How many planets are there in the Solar System?
    ('In',)
    ('In the Solar System, there are **eight planets**: Mercury, Venus, Earth, Mars, Jupiter, Saturn',)
    ('In the Solar System, there are **eight planets**: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune.',)
    ('In the Solar System, there are **eight planets**: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune.',)
    run event=RunStreamEvent(status=<StreamEvent.DONE: 3>, error=None, _message=Message
    (id='fvt4f3p6ddue********', parts=('In the Solar System, there are **eight planets**: Mercury, Venus, Earth, 
    Mars, Jupiter, Saturn, Uranus, and Neptune.',), thread_id='fvt50ma5302n********', 
    created_by='ajegtlf2q28a********', created_at=datetime.datetime(2025, 3, 13, 17, 51, 33, 36643), 
    labels=None, author=Author(id='fvtle31p6lv2********', role='ASSISTANT'), citations=()))
    Ask your question to the assistant (or "exit" to exit): How far is it from the Sun to the Earth?
    ('The average',)
    ('The average distance from the Earth to the Sun is approximately 149.6 million kilometers. This 
    distance',)
    ('The average distance from the Earth to the Sun is approximately 149.6 million kilometers. This 
    distance is also known as the astronomical unit (AU).',)
    ('The average distance from the Earth to the Sun is approximately 149.6 million kilometers. This 
    distance is also known as the astronomical unit (AU).',)
    run event=RunStreamEvent(status=<StreamEvent.DONE: 3>, error=None, _message=Message
    (id='fvtees4295mr********', parts=('The average distance from the Earth to the Sun is approximately 149.
    6 million kilometers. This distance is also known as the astronomical unit (AU).',), 
    thread_id='fvt50ma5302n********', created_by='ajegtlf2q28a********', created_at=datetime.datetime(2025, 
    3, 13, 17, 51, 44, 33797), labels=None, author=Author(id='fvtle31p6lv2********', role='ASSISTANT'), 
    citations=()))
    Ask your question to the assistant (or "exit" to exit): Exit
    Outputting the whole message history when exiting the chat:
        message=Message(id='fvtees4295mr********', parts=('The average distance from the Earth to the Sun is 
        approximately 149.6 million kilometers. This distance is also known as the astronomical unit 
        (AU).',), thread_id='fvt50ma5302n********', created_by='ajegtlf2q28a********', created_at=datetime.
        datetime(2025, 3, 13, 17, 51, 44, 33798), labels=None, author=Author(id='fvtle31p6lv2********', 
        role='ASSISTANT'), citations=())
        message.text='The average distance from the Earth to the Sun is approximately 149.6 million 
        kilometers. This distance is also known as the astronomical unit (AU).'
    
        message=Message(id='fvto6b4rdg0o********', parts=('How far is it from the Sun to the Earth?',), 
        thread_id='fvt50ma5302n********', created_by='ajegtlf2q28a********', created_at=datetime.datetime
        (2025, 3, 13, 17, 51, 42, 941742), labels=None, author=Author(id='fvtjnthkl0g5********', 
        role='USER'), citations=())
        message.text='How far is it from the Sun to the Earth?'
    
        message=Message(id='fvt4f3p6ddue********', parts=('In the Solar System, there are **eight planets**: Mercury, 
        Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune.',), thread_id='fvt50ma5302n********', 
        created_by='ajegtlf2q28a********', created_at=datetime.datetime(2025, 3, 13, 17, 51, 33, 36644), 
        labels=None, author=Author(id='fvtle31p6lv2********', role='ASSISTANT'), citations=())
        message.text='In the Solar System, there are **eight planets**: Mercury, Venus, Earth, Mars, Jupiter, Saturn, 
        Uranus, and Neptune.'
    
        message=Message(id='fvtme86dsuju********', parts=('How many planets are there in the Solar System?' 
        thread_id='fvt50ma5302n********', created_by='ajegtlf2q28a********', created_at=datetime.datetime
        (2025, 3, 13, 17, 51, 32, 139), labels=None, author=Author(id='fvtjnthkl0g5********', role='USER'), 
        citations=())
        message.text='How many planets are there in the Solar System?'
    
        message=Message(id='fvtbkt2tbf7a********', parts=('Hello! What can I do for you?',), 
        thread_id='fvt50ma5302n********', created_by='ajegtlf2q28a********', created_at=datetime.datetime
        (2025, 3, 13, 17, 51, 22, 146834), labels=None, author=Author(id='fvtle31p6lv2********', 
        role='ASSISTANT'), citations=())
        message.text='Hello! What can I do for you?'
    
        message=Message(id='fvtemh1qqc50********', parts=('Hi!',), thread_id='fvt50ma5302n********', 
        created_by='ajegtlf2q28a********', created_at=datetime.datetime(2025, 3, 13, 17, 51, 21, 359885), 
        labels=None, author=Author(id='fvtjnthkl0g5********', role='USER'), citations=())
        message.text='Hi!'
    

See also

  • Creating a simple assistant
  • Creating an assistant with a search index
  • Examples of working with ML SDK on GitHub

Was the article helpful?

Previous
Creating a search assistant specifying file metadata and indexes
Next
About Yandex Foundation Models
© 2025 Direct Cursus Technology L.L.C.