Switching from the AI Assistant API to Responses API

from openai import OpenAI

YANDEX_CLOUD_FOLDER = "<folder_ID>"
YANDEX_CLOUD_MODEL = "<model_URI>"
YANDEX_CLOUD_API_KEY = "<service_account_API_key>"
# or YANDEX_CLOUD_IAM_TOKEN = "<IAM_token>"

previous_id = None  # saving the ID of the last assistant response

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

print("Agent chat (to exit, type ‘exit')\n")

while True:
    user_input = input("You: ")
    if user_input.lower() in ("exit", "quit"):
        print("Chat session ended.")
        break

    response = client.responses.create(
        model=f"gpt://{YANDEX_CLOUD_FOLDER}/{YANDEX_CLOUD_MODEL}",
        input=[{"role": "user", "content": user_input}],
        instructions="You are a text agent that maintains a conversation and provides meaningful responses to the user’s questions.",
        previous_response_id=previous_id,  # providing context, if any
    )

    # saving the ID for the next step
    previous_id = response.id

    # outputting the agent's response
    print("Agent:", response.output_text)

# First, a tool is created to work with the existing search index.
tool = sdk.tools.search_index(
    search_index,
    call_strategy={
        "type": "function",
        "function": {"name": "guide", "instruction": instruction},
    },
)

# Next, an assistant is created which is going to use that tool.
assistant = sdk.assistants.create(
    "yandexgpt",
    instruction="You are an internal corporate documentation assistant. Answer politely. If the information is not in the documents below, don't make up your answer.",
    tools=[tool],
)
thread = sdk.threads.create()

import openai
import json

YANDEX_CLOUD_FOLDER = "<folder_ID>"
YANDEX_CLOUD_MODEL = "<model_URI>"
VECTOR_STORE_ID = "<Vector_Store_instance_ID>"
YANDEX_CLOUD_API_KEY = "<service_account_API_key>"
# or YANDEX_CLOUD_IAM_TOKEN = "<IAM_token>"


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

response = client.responses.create(
    model=f"gpt://{YANDEX_CLOUD_FOLDER}/{YANDEX_CLOUD_MODEL}",
    instructions="You are a smart assistant. If asked about ..., search through the index for information",
    tools=[
        {
            "type": "file_search", 
            "vector_store_ids": [VECTOR_STORE_ID],
        }
    ],
    input="what is ...",
)

print("Response text:")
print(response.output_text)
print("\n" + "=" * 50 + "\n")

# Full response
print("Full response (JSON):")
print(json.dumps(response.model_dump(), indent=2, ensure_ascii=False))

{
  "folderId": "<folder_ID>",
  "modelUri": "gpt://<folder_ID>/yandexgpt-lite/latest",
  "instruction": "You are a smart assistant designed for a finance company. Answer politely. Use search to answer the questions. Do not make up your answer.",
  "tools": [
    {
      "genSearch": {
        "options": {
          "site": {
            "site": [
              "https://cbr.ru/",
              "https://yandex.ru/finance/currencies"
            ]
          },
          "enableNrfmDocs": true
        },
        "description": "Tool to get information about official currency exchange rates."
      }
    }
  ]
}

import openai
import json

YANDEX_CLOUD_FOLDER = "<folder_ID>"
YANDEX_CLOUD_MODEL = "<model_URI>"
YANDEX_CLOUD_API_KEY = "<service_account_API_key>"
# or YANDEX_CLOUD_IAM_TOKEN = "<IAM_token>"


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

response = client.responses.create(
    model=f"gpt://{YANDEX_CLOUD_FOLDER}/{YANDEX_CLOUD_MODEL}",
    input="Create a brief overview of the latest LLM news in 2025, only facts, no specualtion.",
    # Providing tool setting
    tools=[
        {
            "type": "web_search",
            "filters": {
                "allowed_domains": [
                    "habr.ru",
                ],
                "user_location": {
                    "region": "213",
                }
            }
        }
      ],
    temperature=0.3,
    max_output_tokens=1000,
)

run = assistant.run_stream(thread)

# Intermediate results you are getting as the model is generating its response
for event in run:
    print(event._message.parts)

# All fields of the final result
print(f"run {event=}")

import openai

YANDEX_CLOUD_FOLDER = "<folder_ID>"
YANDEX_CLOUD_MODEL = "<model_URI>"
YANDEX_CLOUD_API_KEY = "<service_account_API_key>"
# or YANDEX_CLOUD_IAM_TOKEN = "<IAM_token>"

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

# Creating a streaming request
with client.responses.stream(
    model=f"gpt://{YANDEX_CLOUD_FOLDER}/{YANDEX_CLOUD_MODEL}",
    input="Write a short friendly and funny birthday toast.",
) as stream:
    for event in stream:
        # Text response deltas
        if event.type == "response.output_text.delta":
            print(event.delta, end="", flush=True)
        # Event showing that the response is completed
        # elif event.type == "response.completed":
        #     print("\n---\nResponse completed")

    # You can get the full text of the response if you need to
    # final_response = stream.get_final_response()
    # print("\nFull response text:\n", final_response.output_text)