Contact UsTry it for free
© 2026 Direct Cursus Technology L.L.C.

Creating a simple text agent with a function call

Written by
Updated at December 23, 2025

In Yandex AI Studio, you can create a text agent with a function call that can communicate with the user in text format and maintain a dialogue closely resembling natural human interaction.

Getting started

To use an example:

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

  2. Get and save the service account's API key with yc.ai.foundationModels.execute for its scope.

    Note

    If you are using Windows, we recommend installing the WSL shell first and using it to proceed.

  3. Install Python 3.10 or higher.

  4. Install Python venv to create isolated virtual environments in Python.

  5. Create a new Python virtual environment and activate it:

    python3 -m venv new-env
source new-env/bin/activate

  6. Install the OpenAI library for Python:

    pip install openai

Build a chat

  1. Create a file named index.py and add the following code to it:

    import openai
from openai import OpenAI
import json

YANDEX_CLOUD_MODEL = "yandexgpt"

client = openai.OpenAI(
    api_key=YANDEX_CLOUD_API_KEY,
    base_url="https://rest-assistant.api.cloud.yandex.net/v1",
    project=YANDEX_CLOUD_FOLDER
)

# 1. Listing the functions the model can call
tools = [
    {
        "type": "function",
        "name": "get_weather",
        "description": "Getting current weather for the city you specify.",
        "parameters": {
            "type": "object",
            "properties": {
                "city": {
                    "type": "string",
                    "description": "City, e.g., Petersburg or Moscow",
                },
            },
            "required": ["city"],
        },
    },
]

# The simplest implementation of the function (can be replaced with a call to a real API)
def get_weather(city):
    # Here you can implement an integration, e.g., with Yandex Weather
    return {
        "city": city,
        "temperature": "12 °C",
        "condition": "Cloudy, light breeze"
    }

# Forming a list of messages to be updated
input_list = [
    {"role": "user", "content": "What is the weather like in Krasnoyarsk?"}
]

# 2. Requesting a model with specific features
response = client.responses.create(
    model=f"gpt://{YANDEX_CLOUD_FOLDER}/{YANDEX_CLOUD_MODEL}",
    tools=tools,
    input=input_list,
)

# Adding the model's input to history
input_list += response.output

for item in response.output:
    if item.type == "function_call":
        if item.name == "get_weather":
            # 3. Executing the get_weather function
            weather_info = get_weather(**json.loads(item.arguments))

            # 4. Providing the result of the function back to the model
            input_list.append({
                "type": "function_call_output",
                "call_id": item.call_id,
                "output": json.dumps(weather_info)
            })

print("Final input:")
for item in input_list:
    if isinstance(item, dict) and item.get("type") == "function_call_output":
        parsed = json.loads(item["output"])
        print("function_call_output:", parsed)
    else:
        print(item)

response = client.responses.create(
    model=f"gpt://{YANDEX_CLOUD_FOLDER}/{YANDEX_CLOUD_MODEL}",
    instructions="Respond only with the weather data returned by the function.",
    tools=tools,
    input=input_list,
)

# 5. Final response
print("Final output:")
print(response.model_dump_json(indent=2))
print("\n" + response.output_text)

  2. Save authentication data into environment variables:

    export YANDEX_CLOUD_FOLDER=<folder_ID>
export YANDEX_CLOUD_API_KEY=<API_key>

  3. Run the file you created:

    python index.py

    Response example:

    Final input:
{'role': 'user', 'content': 'What is the weather like in Krasnoyarsk?'}
ResponseFunctionToolCall(arguments='{"city":"Krasnoyarsk"}', call_id='get_weather', name='get_weather', type='function_call', id='get_weather', status='completed', valid=True)
function_call_output: {'city': 'Krasnoyarsk', 'temperature': '12 °C', 'condition': 'Cloudy, light breeze'}
Final output:
{
  "id": "70d96fac-1c4b-4f4a-9f80-56df********",
  "created_at": 1758556157206.0,
  "error": null,
  "incomplete_details": null,
  "instructions": "Respond only with the weather data returned by the function.",
  "metadata": null,
  "model": "gpt://b1gstllj8rgs********/yandexgpt",
  "object": "response",
  "output": [
    {
      "id": "f15c66e8-99a2-4647-a820-406e********",
      "content": [
        {
          "annotations": [],
          "text": "It is currently plus 12°C in Krasnoyarsk, cloudy, light breeze.",
          "type": "output_text",
          "logprobs": null,
          "valid": true
        }
      ],
      "role": "assistant",
      "status": "completed",
      "type": "message",
      "valid": true
    }
  ],
  "parallel_tool_calls": true,
  "temperature": null,
  "tool_choice": "auto",
  "tools": [
    {
      "name": "get_weather",
      "parameters": {
        "type": "object",
        "properties": {
          "city": {
            "type": "string",
            "description": "City, e.g., Petersburg or Moscow"
          }
        },
        "required": [
          "city"
        ]
      },
      "strict": null,
      "type": "function",
      "description": "Getting current weather for the city you specify.",
      "valid": true
    }
  ],
  "top_p": null,
  "background": false,
  "conversation": null,
  "max_output_tokens": null,
  "max_tool_calls": null,
  "previous_response_id": null,
  "prompt": null,
  "prompt_cache_key": null,
  "reasoning": null,
  "safety_identifier": null,
  "service_tier": null,
  "status": "completed",
  "text": null,
  "top_logprobs": null,
  "truncation": null,
  "usage": null,
  "user": "",
  "valid": true
}

It is currently plus 12°C in Krasnoyarsk, cloudy, light breeze.

See also

Previous
Creating a simple text agent
Next
Creating a voice agent via Realtime API