### Install SDK with All Extras

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Install the SDK with all optional extras for maximum compatibility.

```bash
pip install "yandex-ai-studio-sdk[cli-wiki,cli-s3]"
```

--------------------------------

### Install Documentation Dependencies

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/CONTRIBUTING.md

Install Sphinx and other necessary packages for working with project documentation. This command should be run from the repository root.

```bash
pip install -r docs/requirements.txt -e .
```

--------------------------------

### Logging Setup

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/sdk.md

Configures the default logging for the SDK.

```APIDOC
## setup_default_logging(log_level='INFO', log_format='[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s', date_format='%Y-%m-%d %H:%M:%S')

### Description
Sets up the default logging configuration for the SDK. Allows customization of log level, format, and date format.

### Parameters
- **log_level** (string) - Optional - The logging level (e.g., 'INFO', 'DEBUG'). Defaults to 'INFO'.
- **log_format** (string) - Optional - The format string for log messages. Defaults to '[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s'.
- **date_format** (string) - Optional - The date format for log timestamps. Defaults to '%Y-%m-%d %H:%M:%S'.
```

--------------------------------

### Install Vector Stores CLI

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Commands to install the Yandex AI Studio SDK and its optional components for S3 and Wikipedia support.

```bash
pip install yandex-ai-studio-sdk
pip install "yandex-ai-studio-sdk[cli-s3]"    # S3 / Yandex Object Storage support
pip install "yandex-ai-studio-sdk[cli-wiki]"   # Wikipedia / MediaWiki support
```

--------------------------------

### Integer Bit Length Example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/domain.md

Demonstrates how to get the number of bits necessary to represent an integer in binary.

```python
>>> bin(37)
'0b100101'
>>> (37).bit_length()
6
```

--------------------------------

### Install CLI with Wikipedia/MediaWiki Support for Yandex AI Studio SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the Yandex AI Studio SDK CLI with support for Wikipedia and MediaWiki. This allows indexing content from these platforms.

```sh
pip install "yandex-ai-studio-sdk[cli-wiki]"
```

--------------------------------

### Install Yandex AI Studio SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the SDK using pip. This command is used to add the library to your Python environment.

```sh
pip install yandex-ai-studio-sdk
```

--------------------------------

### Install Test Dependencies

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/CONTRIBUTING.md

Install all necessary dependencies for testing, including extra requirements, directly using pip. This bypasses tox for dependency management.

```bash
pip install -r test_requirements.txt -r test_requirements_extra.txt
```

--------------------------------

### Install CLI with S3 Support for Yandex AI Studio SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the Yandex AI Studio SDK CLI with support for S3-compatible storage, including Yandex Object Storage. Use this if you need to index data from S3.

```sh
pip install "yandex-ai-studio-sdk[cli-s3]"
```

--------------------------------

### Install LangChain Extra for Yandex AI Studio SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the SDK with the 'langchain' extra to enable LangChain integration. This is required before using LangChain features.

```sh
pip install yandex-ai-studio-sdk[langchain]
```

--------------------------------

### Async Streaming Example with YandexGPT

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Demonstrates how to use the SDK for asynchronous streaming of responses from a YandexGPT model. Requires importing asyncio and the AIStudio SDK.

```python
import asyncio
from langchain_core.messages import HumanMessage

from yandex.ai.studio.sdk import AIStudio

async def stream_example():
    sdk_async = AIStudio()
    alc_model = sdk_async.models.completions('yandexgpt').langchain(model_type="chat")
    async for chunk in alc_model.astream([HumanMessage(content="Tell me a joke")]):
        print(chunk.content, end="", flush=True)

asyncio.run(stream_example())
```

--------------------------------

### Integer to Ratio Example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/domain.md

Illustrates converting an integer into a pair of integers representing its ratio with a positive denominator.

```python
>>> (10).as_integer_ratio()
(10, 1)
>>> (-10).as_integer_ratio()
(-10, 1)
>>> (0).as_integer_ratio()
(0, 1)
```

--------------------------------

### Text Output Example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Displays the success message and index details in plain text format.

```text
Search index created successfully!
Search Index ID: fvt-xxxxxxxxxxxxxxxx
Name: my-index
```

--------------------------------

### RecognitionClassifier Initialization

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.md

Demonstrates how to initialize a `RecognitionClassifier` with a string trigger and trigger type. Shows examples with single and multiple triggers, and using the `on_utterance` alternative constructor.

```python
RecognitionClassifier('insult', 'on_final')
```

```python
RecognitionClassifier('insult', ['on_final', 'on_partial'])
```

```python
RecognitionClassifier.on_utterance('insult')
```

--------------------------------

### Index Multiple Local Files

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Example of indexing multiple individual text and markdown files.

```bash
yandex-ai-studio vector-stores local docs/intro.txt docs/guide.md
```

--------------------------------

### Index Local File with Custom Name

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Example of indexing a local file and specifying a custom name for the search index.

```bash
yandex-ai-studio vector-stores local report.pdf --name "Q4 Report"
```

--------------------------------

### Simple chat completion run

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Shows a basic example of running a chat completion with a specified model and temperature.

```APIDOC
## Simple run

### Description
Performs a simple chat completion request using a specified model and configures the temperature.

### Method
`model.run(prompt)`

### Parameters
#### Model Configuration
- **temperature** (float) - Optional - Controls the randomness of the output.

### Request Example
```python
from yandex_ai_studio_sdk import AIStudio

sdk = AIStudio()
model = sdk.chat.completions('qwen3-235b-a22b-fp8').configure(temperature=0.5)
result = model.run("How to calculate the Hirsch index in O(N)?")
print(result.text)
```
```

--------------------------------

### JSON Output Example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Displays the success message and index details in JSON format when the --format json option is used.

```json
{
  "status": "success",
  "folder_id": "b1gxxxxxxxxxxxxx",
  "search_index": {
    "id": "fvt-xxxxxxxxxxxxxxxx",
    "name": "my-index"
  }
}
```

--------------------------------

### Index Local Files with Shell Globbing

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Example using shell globbing to include all .txt and .md files from a directory.

```bash
yandex-ai-studio vector-stores local sample_docs/*.txt sample_docs/*.md
```

--------------------------------

### Configure Web Search Settings

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_api.md

Use the `.configure` method to change initial search settings for web searches. This example shows how to switch the search type from 'BE' to 'RU'.

```python
>>> search = sdk.search_api.web(search_type='BE')
>>> search = search.configure(search_type='RU')
```

--------------------------------

### Basic SDK Usage for Text Generation

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Demonstrates initializing the SDK with folder ID and authentication, configuring a completion model, and running a text generation request. Ensure you replace placeholders with your actual credentials and folder ID.

```python
from yandex_ai_studio_sdk import AIStudio


sdk = AIStudio(folder_id="...", auth="<APIKey/IAMToken/SomethingElse>")

model = sdk.models.completions('yandexgpt')
model = model.configure(temperature=0.5)
result = model.run("foo")

for alternative in result:
    print(alternative)
```

--------------------------------

### Compute Text Embeddings and Similarities

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Use 'query' and 'doc' model variants to get vector representations of text and compute cosine similarity. Ensure `scipy` and `numpy` are installed.

```python
import numpy as np
from scipy.spatial.distance import cdist
from yandex_ai_studio_sdk import AIStudio

 sdk = AIStudio()

 doc_texts = [
    "Alexander Pushkin (1799–1837) was a Russian poet and novelist.",
    "Chamomile is a flowering plant of the family Asteraceae.",
 ]

 query_model = sdk.models.text_embeddings('query')
 doc_model = sdk.models.text_embeddings('doc')

 query_result = query_model.run("When was Pushkin born?")
 doc_results = [doc_model.run(text) for text in doc_texts]

 query_vec = np.array(query_result.embedding)
 doc_vecs = np.array([r.embedding for r in doc_results])

 similarities = 1 - cdist([query_vec], doc_vecs, metric='cosine')[0]
 best_match = doc_texts[np.argmax(similarities)]
 print(best_match)
# → "Alexander Pushkin (1799–1837) was a Russian poet and novelist."
```

--------------------------------

### Serve Documentation with Auto-rebuild

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/CONTRIBUTING.md

Run a local web server to view the generated documentation and automatically rebuild it when source files change. This provides a live preview during development.

```bash
sphinx-autobuild docs docs/_build --watch src
```

--------------------------------

### Get Binary Representation of an Integer

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/batch.md

Demonstrates how to get the binary representation of an integer using the bin() function.

```python
>>> bin(37)
'0b100101'
```

--------------------------------

### AsyncAIStudio Initialization

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.md

Instantiate the main class to interact with the SDK. Various optional parameters can be provided for configuration.

```APIDOC
## AsyncAIStudio Constructor

### Description
Construct a new asynchronous SDK instance.

### Parameters
* **folder_id** (str) - Optional - Yandex Cloud folder identifier.
* **endpoint** (str) - Optional - Domain and port for the Yandex Cloud API.
* **auth** (str | BaseAuth | Undefined) - Optional - Authentication credentials.
* **service_map** (Dict[str, str]) - Optional - Dictionary to redefine service endpoints.
* **enable_server_data_logging** (bool | Undefined) - Optional - Enable or disable server-side data logging.
* **verify** (bool | pathlib.Path | str | os.PathLike) - Optional - SSL certificates for verification.
* **retry_policy** (RetryPolicy | Undefined) - Optional - Retry policy configuration.
* **yc_profile** (str | Undefined) - Optional - Yandex Cloud profile name.
* **interceptors** (Sequence[ClientInterceptor] | Undefined) - Optional - Sequence of client interceptors.
```

--------------------------------

### Configure Text-to-Speech Settings

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/text_to_speech.md

Demonstrates how to initialize and reconfigure Text-to-Speech settings, such as audio format. The `configure` method allows overriding initial settings.

```python
>>> tts = sdk.speechkit.text_to_speech(audio_format='mp3')
>>> tts = tts.configure(audio_format='WAV')
```

--------------------------------

### Confluence On-Premise URL Example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Example of an on-premise Confluence page URL format that includes a numeric page ID.

```bash
https://confluence.example.com/pages/viewpage.action?pageId=123456
```

--------------------------------

### Confluence Cloud URL Example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Example of a Confluence Cloud page URL format that includes a numeric page ID.

```bash
https://your-domain.atlassian.net/wiki/spaces/SPACE/pages/123456/Page+Title
```

--------------------------------

### setup_default_logging

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/internals/bases.md

Sets up the default logging configuration for the SDK.

```APIDOC
## setup_default_logging(log_level='INFO', log_format='[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s', date_format='%Y-%m-%d %H:%M:%S')

Sets up the default logging configuration.

Read more about log_levels, log_format, and date_format in [Python documentation (logging)](https://docs.python.org/3/library/logging.html).

### Parameters
* **log_level** (Literal['CRITICAL', 'FATAL', 'ERROR', 'WARN', 'WARNING', 'INFO', 'DEBUG', 'TRACE'] | ~typing.Literal['critical', 'fatal', 'error', 'warn', 'warning', 'info', 'debug', 'TRACE'] | int) – The logging level to set.
* **log_format** (str) – The format of the log messages.
* **date_format** (str) – The format for timestamps in log messages.

### Returns
The instance of the SDK with logging configured.

### Return type
Self
```

--------------------------------

### setup_default_logging

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.md

Sets up the default logging configuration for the SDK. Allows customization of log level, format, and date format.

```APIDOC
## setup_default_logging(log_level='INFO', log_format='[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s', date_format='%Y-%m-%d %H:%M:%S')

### Description
Sets up the default logging configuration for the SDK. This function allows users to configure the verbosity and format of log messages generated by the SDK.

### Parameters
#### Keyword Parameters
* **log_level** (Literal['CRITICAL', 'FATAL', 'ERROR', 'WARN', 'WARNING', 'INFO', 'DEBUG', 'TRACE'] | ~typing.Literal['critical', 'fatal', 'error', 'warn', 'warning', 'info', 'debug', 'TRACE'] | int) - Optional - The logging level to set. Defaults to 'INFO'.
* **log_format** (str) - Optional - The format of the log messages. Defaults to '[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s'.
* **date_format** (str) - Optional - The format for timestamps in log messages. Defaults to '%Y-%m-%d %H:%M:%S'.

### Returns
The instance of the SDK with logging configured.

### Return type
Self
```

--------------------------------

### Create and Manage Search Indexes

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Demonstrates creating a text search index, adding files to it, renaming, and listing files. Ensure files are uploaded before creating an index.

```python
from yandex_ai_studio_sdk import AIStudio
from yandex_ai_studio_sdk.search_indexes import (
    StaticIndexChunkingStrategy,
    TextSearchIndexType,
    VectorSearchIndexType,
    HybridSearchIndexType,
)

sdk = AIStudio()

# Upload source files
files = [
    sdk.files.upload('docs/turkey.txt', ttl_days=5, expiration_policy='static'),
    sdk.files.upload('docs/maldives.txt', ttl_days=5, expiration_policy='static'),
]

# Create text (BM25) search index
op = sdk.search_indexes.create_deferred(
    [files[0]],
    index_type=TextSearchIndexType(
        chunking_strategy=StaticIndexChunkingStrategy(
            max_chunk_size_tokens=700,
            chunk_overlap_tokens=300,
        )
    ),
)
index = op.wait()
print(f"Index created: {index.id}")

# Add more files to an existing index
add_op = index.add_files_deferred(files[1])
add_op.wait()

# Rename the index
index.update(name="travel-knowledge-base")

# List files in index
for f in index.list_files():
    print(f.id)

# List and delete all indexes
for idx in sdk.search_indexes.list():
    idx.delete()
```

--------------------------------

### LLMPostProcessing Class Initialization

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.md

Demonstrates how to initialize the LLMPostProcessing class with a model URI, model name, or model name and version.

```APIDOC
## LLMPostProcessing Initialization

### Description
Initializes the LLMPostProcessing class to configure post-processing for speech-to-text results using a specified language model.

### Parameters
* **model_name** (str) - The identifier for the language model. Can be a full URI (e.g., 'gpt://<folder_id>/yandexgt/latest') or a model name (e.g., 'yandexgt'). If a model name is provided without a URI scheme, it will be constructed using the configured folder ID.
* **model_version** (str, optional) - The version of the model to use. Defaults to 'latest'.
* **instructions** (Sequence[LLMPostProcessingInstruction], optional) - An initial sequence of instructions to apply. Defaults to an empty tuple.

### Request Example
```python
# Using a full model URI
llm_post_processor_uri = LLMPostProcessing('gpt://<folder_id>/yandexgt/latest')

# Using a model name (folder_id is configured separately)
llm_post_processor_name = LLMPostProcessing('yandexgt')

# Using a model name and specific version
llm_post_processor_version = LLMPostProcessing('yandexgt', model_version='2024-04-09')
```
```

--------------------------------

### Initialize AsyncAIStudio SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/examples/async/image_generation/run_deferred.ipynb

Initializes the asynchronous SDK client. Authentication can be managed via environment variables for tokens and folder IDs.

```python
from yandex_ai_studio_sdk import AsyncAIStudio
# You can set authentication using environment variables instead of the 'auth' argument:
# YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, or YC_API_KEY
# You can also set 'folder_id' using the YC_FOLDER_ID environment variable
sdk = AsyncAIStudio(
    # folder_id="<YC_FOLDER_ID>",
    # auth="<YC_API_KEY/YC_IAM_TOKEN>",
)
model = sdk.models.image_generation('yandex-art')
```

--------------------------------

### Adding Instructions with LLMPostProcessing

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.md

Shows how to add instructions to an LLMPostProcessing object using the `with_instruction` method.

```APIDOC
## LLMPostProcessing with_instruction Method

### Description
Appends a new instruction to the LLMPostProcessing object, returning a new instance with the added instruction. This method does not mutate the original object.

### Parameters
* **instruction** (str) - The instruction text to be given to the language model.
* **response_format** (Literal['json'] | ~yandex_ai_studio_sdk._types.schemas.JsonSchemaResponseType | type | None, optional) - Specifies the desired format of the model's response. 
    - `None`: Returns a standard text response (default).
    - `'json'`: Instructs the model to return a JSON object.
    - `dict`: A dictionary representing a JSON schema to enforce a specific JSON structure.
    - `pydantic model or pydantic dataclass`: The SDK will transform this into a JSON schema.

### Return type
LLMPostProcessing - A new LLMPostProcessing object with the appended instruction.

### Request Example
```python
# Initialize post processor
llm_post_processor = LLMPostProcessing('yandexgt')

# Add a simple text instruction
llm_post_processor = llm_post_processor.with_instruction("Make a short review")

# Add an instruction with JSON response format
llm_post_processor = llm_post_processor.with_instruction(
    "What the conversation topic",
    response_format="json"
)

# Add an instruction with a specific JSON schema response format
# Assuming MySchema is a defined Pydantic model or dict schema
# llm_post_processor = llm_post_processor.with_instruction(
#     "Extract key entities",
#     response_format=MySchema
# )
```

### Object Representation After Adding Instructions
```python
# Example of how the object might look after adding instructions
# LLMPostProcessing(model_name='yandexgt', model_version='latest', ..., instructions=(LLMPostProcessingInstruction(...), LLMPostProcessingInstruction(...)))
```
```

--------------------------------

### id

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/tuning.md

Gets the unique identifier for the fine-tuning task.

```APIDOC
## id

### Description
Get fine-tuning task identifier.

### Return type
str
```

--------------------------------

### Manage Datasets for Fine-tuning and Batch Inference

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Demonstrates creating, uploading, and managing datasets. Supports both a shortcut API for completions datasets and a generic API for other task types. Datasets can be uploaded directly or deferred.

```python
from yandex_ai_studio_sdk import AIStudio

 sdk = AIStudio()

# --- Completions dataset (shortcut API) ---
draft = sdk.datasets.completions.draft_from_path(
    'training_data.jsonlines',
    upload_format='jsonlines',
    name='my-completions-v1',
)
draft.allow_data_logging = True
dataset = draft.upload()
print(f"Dataset ready: {dataset.id}, status={dataset.status}")

# --- Generic dataset ---
draft2 = sdk.datasets.draft_from_path(
    task_type='TextToTextGeneration',
    path='data.jsonlines',
    upload_format='jsonlines',
    name='generic-dataset',
)
# Deferred upload: returns Operation to track progress
op = draft2.upload_deferred()
dataset2 = op.wait()

# List datasets with filters
for ds in sdk.datasets.completions.list(name_pattern='my-completions'):
    print(ds.id, ds.status)

# Get single dataset by ID
ds = sdk.datasets.get('some-dataset-id')
ds.delete()
```

--------------------------------

### Initialize AI Studio SDK

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Instantiate the SDK client for synchronous or asynchronous operations. Authentication and folder ID can be provided explicitly or inferred from environment variables.

```python
from yandex_ai_studio_sdk import AIStudio, AsyncAIStudio

# Sync client — auth and folder_id read from env vars
sdk = AIStudio().setup_default_logging()

# Explicit credentials
 sdk = AIStudio(
 folder_id="b1gxxxxxxxxxxxxxxxxx",
 auth="AQVNxxxxxxxxxxxxxxxxx",  # API key, IAM token, or OAuth token
)

# Async client for use inside asyncio applications
import asyncio
from yandex_ai_studio_sdk import AsyncAIStudio

async def main():
 sdk = AsyncAIStudio(folder_id="b1gxxxxxxxxxxxxxxxxx", auth="AQVNxxxxxxxxxxxxxxxxx")
 model = sdk.models.completions('yandexgpt')
 result = await model.configure(temperature=0.5).run("Hello!")
 print(result[0].text)

asyncio.run(main())
```

--------------------------------

### Get Thread

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/threads.md

Retrieves a specific thread by its unique identifier.

```APIDOC
## get

### Description
Retrieve a thread by its id.

This method fetches an already created thread using its unique identifier.

### Method Signature
`get(thread_id, timeout=60)`

### Parameters
#### Path Parameters
- **thread_id** (str) - the unique identifier of the thread to retrieve.
#### Optional Parameters
- **timeout** (float) - timeout for the service call in seconds. Defaults to 60 seconds.

### Return Type
`Thread`
```

--------------------------------

### Get Assistant

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/assistants.md

Retrieves an existing assistant by its unique ID.

```APIDOC
## get Assistant

### Description
Get an existing assistant by ID.

### Method
`get`

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **assistant_id** (str) - Required - ID of the assistant to retrieve
- **timeout** (float) - Optional - Timeout in seconds (defaults to 60)

### Request Example
```python
assistant = assistants.get("assistant-id")
```

### Response
#### Success Response (200)
- **Assistant** (Assistant) - The retrieved assistant object
```

--------------------------------

### Get Search Index

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/search_indexes.md

Retrieves a search index by its unique identifier.

```APIDOC
## get

### Description
Retrieve a search index by its id. This method fetches an already created search index using its unique identifier.

### Method Signature
`get(search_index_id, timeout=60)`

### Parameters
* **search_index_id** (*str*) – Required - The unique identifier of the search index to retrieve.
* **timeout** (*float*) – Optional - The time to wait for the operation to complete. Defaults to 60 seconds.

### Return type
*SearchIndex*
```

--------------------------------

### Initialize AsyncAIStudio SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/examples/async/speechkit/text_to_speech/simple_run.ipynb

Initializes the asynchronous AI Studio SDK. Authentication can be set via environment variables or directly using the 'auth' argument. The folder ID can also be configured.

```python
from yandex_ai_studio_sdk import AsyncAIStudio
from yandex_ai_studio_sdk._experimental.audio.out import AsyncAudioOut

SAMPLERATE = 44100
```

```python
# You can set authentication using environment variables instead of the 'auth' argument:
# YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, or YC_API_KEY
# You can also set 'folder_id' using the YC_FOLDER_ID environment variable
sdk = AsyncAIStudio(
    # folder_id="<YC_FOLDER_ID>",
    # auth="<YC_API_KEY/YC_IAM_TOKEN>",
).setup_default_logging()
```

--------------------------------

### SDK Initialization — AIStudio / AsyncAIStudio

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Entry point for all SDK functionality. Instantiate once per application, then access feature domains as attributes. Auth and folder ID can come from arguments or environment variables.

```APIDOC
## SDK Initialization — `AIStudio` / `AsyncAIStudio`

Entry point for all SDK functionality. Instantiate once per application, then access feature domains as attributes. Auth and folder ID can come from arguments or environment variables.

```python
from yandex_ai_studio_sdk import AIStudio, AsyncAIStudio

# Sync client — auth and folder_id read from env vars
sdk = AIStudio().setup_default_logging()

# Explicit credentials
sdk = AIStudio(
    folder_id="b1gxxxxxxxxxxxxxxxxx",
    auth="AQVNxxxxxxxxxxxxxxxxx",  # API key, IAM token, or OAuth token
)

# Async client for use inside asyncio applications
import asyncio
from yandex_ai_studio_sdk import AsyncAIStudio

async def main():
    sdk = AsyncAIStudio(folder_id="b1gxxxxxxxxxxxxxxxxx", auth="AQVNxxxxxxxxxxxxxxxxx")
    model = sdk.models.completions('yandexgpt')
    result = await model.configure(temperature=0.5).run("Hello!")
    print(result[0].text)

asyncio.run(main())
```
```

--------------------------------

### Get Assistant

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/assistants.md

Asynchronously retrieves an existing assistant by its unique ID.

```APIDOC
## async get

### Description
Get an existing assistant by ID.

### Method
async

### Parameters
* **assistant_id** (str) - Required - ID of the assistant to retrieve
* **timeout** (float) - Optional - The timeout, or the maximum time to wait for the request to complete in seconds. Defaults to 60 seconds.

### Return type
AsyncAssistant
```

--------------------------------

### Get Dataset

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/datasets.md

Fetches a specific dataset using its unique identifier.

```APIDOC
## get

### Description
Fetch a dataset from the server using its ID.

### Method
async

### Parameters
#### Path Parameters
- **dataset_id** (str) - Required - the unique identifier of the dataset to fetch.
- **timeout** (float) - Optional - the time to wait for the request. Defaults to 60 seconds.

### Return type
AsyncDataset
```

--------------------------------

### Configure ByImage Search

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_api.md

Demonstrates how to configure initial search settings for by_image search. Use the `.configure` method to change parameters like the target site.

```python
>>> search = sdk.search_api.by_image(site="ya.ru")
>>> search = search.configure(site="yandex.ru")
```

--------------------------------

### Get File

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/files.md

Retrieves a file from the server using its unique identifier.

```APIDOC
## async get(file_id, timeout=60)

### Description
Retrieves a file by its ID.

### Parameters
#### Path Parameters
* **file_id** (str) - Required - The unique identifier of the file to retrieve.

#### Query Parameters
* **timeout** (float) - Optional - Timeout for the operation in seconds. Defaults to 60.

### Return type
AsyncFile
```

--------------------------------

### TimeSpan Class

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/speechkit.md

Represents a time interval with a start and end time.

```APIDOC
### Class: yandex_ai_studio_sdk._speechkit.speech_to_text.result.utils.TimeSpan

TimeSpan(start_time_ms: int, end_time_ms: int)

#### Attributes:
* **start_time_ms** (*int*): Audio segment start time in milliseconds.
* **end_time_ms** (*int*): Audio segment end time in milliseconds.

#### Properties:
* **length_ms** (*int*): Returns the length of the given time span in milliseconds.
```

--------------------------------

### Integer Bit Length and Count

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/internals/bases.md

Demonstrates how to get the bit length and bit count of an integer.

```python
>>> bin(37)
'0b100101'
>>> (37).bit_length()
6
```

```python
>>> bin(13)
'0b1101'
>>> (13).bit_count()
3
```

--------------------------------

### Initialize LLMPostProcessing with Full Model URI

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.md

Instantiate the LLMPostProcessing class using a complete model URI. This is useful when you need to specify the exact model and version.

```python
>>> llm_post_processor = LLMPostProcessing('gpt://<folder_id>/yandexgt/latest')
```

--------------------------------

### LLMPostProcessing with_instruction Method

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/speechkit/speech_to_text.md

Shows how to add instructions to an LLMPostProcessing object, optionally specifying a response format.

```APIDOC
## LLMPostProcessing.with_instruction

### Description
Appends a new instruction to the post-processor configuration, returning a new LLMPostProcessing object with the added instruction.

### Parameters
* **instruction** (str) - The instruction text for the language model.
* **response_format** (Literal['json'] | ~yandex_ai_studio_sdk._types.schemas.JsonSchemaResponseType | type | None, optional) - Specifies the desired format of the model's response. Can be None for text, 'json' for JSON output, a JsonSchema, or a Pydantic model/dataclass. Defaults to None.

### Return type
LLMPostProcessing - A new LLMPostProcessing object with the appended instruction.

### Request Example
```py
llm_post_processor = LLMPostProcessing('yandexgt')

# Add a simple text instruction
llm_post_processor = llm_post_processor.with_instruction("Make a short review")

# Add an instruction with JSON response format
llm_post_processor = llm_post_processor.with_instruction(
    "What the conversation topic",
    response_format="json"
)
```
```

--------------------------------

### Initialize SpeechKit SDK

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/text_to_speech.md

Initializes the SpeechKit SDK with necessary parameters. Requires an SDK instance, a URI, and optionally a configuration object and owner.

```python
#### __init__(, sdk, uri, config=None, owner=None)
* **Parameters:**
  * **sdk** ([*yandex_ai_studio_sdk._sdk.BaseSDK*](../../internals/bases.md#yandex_ai_studio_sdk._sdk.BaseSDK))
  * **uri** (*str*)
  * **config** ([*ConfigTypeT](../../types/other.md#yandex_ai_studio_sdk._types.model_config.ConfigTypeT) *|* *None*)
  * **owner** (*str* *|* *None*)
```

--------------------------------

### Run Result Properties

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/runs.md

Access properties of a run result object to get detailed information about the execution.

```APIDOC
## Run Result Object

### Description
Represents the outcome of a run, providing access to text content, status, errors, and tool calls.

### Properties

- **text** (str | None) - The text content of the message if available, otherwise None.
- **status** (StatusTypeT) - The current status of the run.
- **error** (str | None) - An error message if the run failed, otherwise None.
- **tool_calls** (ToolCallList | None) - A list of tool calls made during the run, if any.
```

--------------------------------

### Index MediaWiki Pages with Yandex AI Studio CLI

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Use the `vector-stores wiki` command to create a search index from MediaWiki pages. Requires installation with the `cli-wiki` extra. Supports indexing multiple pages, exporting in different formats (text, html, markdown), and accessing private wikis with credentials.

```bash
yandex-ai-studio vector-stores wiki https://en.wikipedia.org/wiki/Machine_learning
```

```bash
yandex-ai-studio vector-stores wiki \
  https://en.wikipedia.org/wiki/Machine_learning \
  https://en.wikipedia.org/wiki/Neural_network \
  https://en.wikipedia.org/wiki/Transformer_(deep_learning_architecture)
```

```bash
yandex-ai-studio vector-stores wiki \
  "https://en.wikipedia.org/wiki/Python_(programming_language)" \
  --export-format markdown
```

```bash
yandex-ai-studio vector-stores wiki \
  https://wiki.example.com/wiki/Internal_docs \
  --username alice \
  --password secret
```

--------------------------------

### Index a Single Local File

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Example of indexing a single PDF file using the local subcommand.

```bash
yandex-ai-studio vector-stores local report.pdf
```

--------------------------------

### Load Fine-tuned Model by URI

Source: https://context7.com/yandex-cloud/yandex-ai-studio-sdk/llms.txt

Demonstrates how to load and use a fine-tuned model by specifying its unique URI.

```python
# Fine-tuned model by URI
tuned_model = sdk.models.completions("gpt://b1gxxxxx/yandexgpt-lite/some-tuned-uri")
print(tuned_model.run("Hey!")[0].text)
```

--------------------------------

### tune_deferred

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/models/text_embeddings.md

Initiates a deferred tuning process for the model. This method starts a tuning task that runs asynchronously.

```APIDOC
## tune_deferred(train_datasets, embeddings_tune_type, validation_datasets=Undefined, name=Undefined, description=Undefined, labels=Undefined, seed=Undefined, lr=Undefined, n_samples=Undefined, additional_arguments=Undefined, dimensions=Undefined, tuning_type=Undefined, scheduler=Undefined, optimizer=Undefined, timeout=60)

### Description
Initiate a deferred tuning process for the model.

### Parameters
* **train_datasets** (*str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *|* *tuple* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *,* *float* *]*  *|* *Iterable* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *|* *tuple* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *,* *float* *]* *]*  *|* *dict* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *,* *float* *]*) – the dataset objects and/or dataset ids used for training of the model.
* **embeddings_tune_type** (*Literal* *[* *'pair'* *,*  *'triplet'* *]*) – the type of tuning to be applied (e.g. pair or triplet).
* **validation_datasets** (*str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *|* *tuple* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *,* *float* *]*  *|* *Iterable* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *|* *tuple* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *,* *float* *]* *]*  *|* *dict* *[**str* *|* [*BaseDataset*](../../internals/bases.md#yandex_ai_studio_sdk._datasets.dataset.BaseDataset) *,* *float* *]*  *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – the dataset objects and/or dataset ids used for validation of the model.
* **name** (*str* *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – the name of the tuning task.
* **description** (*str* *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – the description of the tuning task.
* **labels** (*dict* *[**str* *,* *str* *]*  *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – labels for the tuning task.
* **seed** (*int* *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – a random seed for reproducibility.
* **lr** (*float* *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – a learning rate for tuning.
* **n_samples** (*int* *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – a number of samples for tuning.
* **additional_arguments** (*str* *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – additional arguments for tuning.
* **dimensions** (*Sequence* *[**int* *]*  *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – dimensions for the embeddings.
* **tuning_type** ([*BaseTuningType*](../../types/tuning.md#yandex_ai_studio_sdk._types.tuning.tuning_types.BaseTuningType) *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – a type of tuning to be applied.
* **scheduler** ([*BaseScheduler*](../../types/tuning.md#yandex_ai_studio_sdk._types.tuning.schedulers.BaseScheduler) *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – a scheduler for tuning.
* **optimizer** ([*BaseOptimizer*](../../types/tuning.md#yandex_ai_studio_sdk._types.tuning.optimizers.BaseOptimizer) *|* [*Undefined*](../../types/other.md#yandex_ai_studio_sdk._types.misc.Undefined)) – an optimizer for tuning.
* **timeout** (*float*) – the timeout, or the maximum time to wait for the request to complete in seconds.
    Defaults to 60 seconds.

### Return type
[*TuningTask*](../../types/tuning.md#yandex_ai_studio_sdk._tuning.tuning_task.TuningTask)[[*TextEmbeddingsModel*](#yandex_ai_studio_sdk._models.text_embeddings.model.TextEmbeddingsModel)]
```

--------------------------------

### Index Data from S3-Compatible Storage with Yandex AI Studio CLI

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Create a search index from an S3-compatible bucket using the `vector-stores s3` command. Requires installation with the `cli-s3` extra. Supports filtering by prefix, custom endpoints, AWS credentials, and include/exclude patterns.

```bash
yandex-ai-studio vector-stores s3 my-bucket
```

```bash
yandex-ai-studio vector-stores s3 my-bucket --prefix docs/
```

```bash
yandex-ai-studio vector-stores s3 my-bucket --include-pattern "*.pdf"
```

```bash
yandex-ai-studio vector-stores s3 my-bucket \
  --endpoint-url https://storage.yandexcloud.net \
  --region-name ru-central1
```

--------------------------------

### Get Recognition Result

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.md

Retrieves the results of an asynchronous speech recognition operation. This method is used after the operation has completed.

```APIDOC
## async get_recognition_result(operation_id, timeout=60)

### Description
Gets results of asynchronous recognition after finishing the operation.

### Parameters
* **operation_id** (str) - Required - The ID of the recognition operation.
* **timeout** (float) - Optional - The maximum time to wait for the results in seconds. Defaults to 60.

### Return type
*AsyncDeferredSpeechToTextResult*
```

--------------------------------

### Get Run Result

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/runs.md

Retrieves the final result of a completed run. This method should be called after the run has finished execution.

```APIDOC
## GET /runs/{run_id}/result

### Description
Get the result of a run.

### Method
GET

### Endpoint
`/runs/{run_id}/result`

### Parameters
#### Path Parameters
- **run_id** (str) - Required - Run ID

#### Query Parameters
- **timeout** (float) - Optional - The timeout, or the maximum time to wait for the request to complete in seconds. Defaults to 60 seconds.

### Response
#### Success Response (200)
- **AnyResultTypeT_co** (*object*) - The final result of the run.

### Response Example
```json
{
  "type": "text",
  "value": "The weather is sunny."
}
```
```

--------------------------------

### Get File from Search Index

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_indexes.md

Retrieves a specific file associated with a search index using its file ID.

```APIDOC
## Get File from Search Index

### Description
Retrieves a file associated with the search index.

### Method
`async get_file`

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
* **file_id** (str) - Required - the ID of the file to retrieve.
* **timeout** (float) - Optional - the time to wait for the get request. Defaults to 60 seconds.

### Request Example
```python
file_data = await search_index.get_file(file_id="file_123", timeout=30.0)
```

### Response
#### Success Response (200)
* **SearchIndexFile** - The retrieved file object.

#### Response Example
```json
{
  "id": "file_123",
  "name": "document.pdf",
  "size": 1024,
  "created_at": "2023-01-01T10:00:00Z"
}
```
```