---
id: gigachat
version: "0.2.3"
license: MIT
license_treatment: permissive
maintenance: active
---

# gigachat — GigaChat. Python-library for GigaChat API

License: permissive · Maintenance: active · Downloads: 143.0K/mo

## What it is and what it does

GigaChat is a Python SDK that wraps the GigaChat REST API, a large language model service. It provides a typed, Pydantic V2-based interface for chat completions, streaming responses, text embeddings, function calling (tool use), and vision capabilities. The library supports both synchronous and asynchronous operations, making it suitable for integration into synchronous scripts, async applications, or frameworks like LangChain through the official langchain-gigachat integration.

The SDK handles authentication through multiple methods (OAuth, password, TLS certificates, access tokens), includes automatic retry logic with configurable exponential backoff, and offers token counting to estimate usage before making requests. It requires Python 3.8 or later and depends on httpx for HTTP operations and pydantic for data validation. The library is actively maintained and marked as production-stable.

Use it for:

- Build chatbots or conversational agents that stream responses in real-time to users.
- Generate vector embeddings for text to power semantic search or similarity-based retrieval systems.
- Create agents that call external functions or tools by leveraging the function-calling capability.
- Integrate GigaChat into LangChain workflows using the official langchain-gigachat integration.
- Analyze images or multimodal content using the vision capabilities of the model.
- Estimate token usage before making API requests to manage costs and request planning.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Python SDK for the GigaChat REST API, a large language model, providing chat completions, streaming, embeddings, function calling, and vision capabilities with synchronous and asynchronous support.

Yes. The package is actively maintained, has low install friction, carries a permissive MIT license, and supports modern Python versions (3.8–3.13). It offers a complete, typed interface to a production-grade LLM service with no known vulnerabilities. Install it if you need to integrate GigaChat into a Python application.

## Install

pip install gigachat
uv add gigachat
poetry add gigachat

## Installing gigachat

Before you install: Low install friction with a pure-wheel distribution and only three runtime dependencies (httpx, pydantic, pydantic-settings). Actively maintained with a recent release 14 days ago and ongoing repository activity.

License in practice: MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects without licensing concerns.

Quickstart:

pip install gigachat

from gigachat import GigaChat

with GigaChat(credentials="<key>", model="GigaChat-2") as client:
    response = client.chat.create("Hello, GigaChat!")
    print(response.messages[0].content[0].text)

Requires GigaChat authorization credentials and a model specification; TLS certificate setup recommended for production use.

Verify before relying:

- Whether the SDK's automatic retry mechanism with exponential backoff is configurable and what the default retry limits are.
- Performance characteristics and rate limits when handling large batches of embeddings or concurrent requests.
- Specific details on which image formats and sizes are supported for vision/multimodal operations.

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 143.0K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

gigachat api python client, large language model sdk, chat completions streaming, text embeddings vectorization, function calling tools agents, async llm integration, multimodal vision api, llm-client, async-support, embeddings

[View on SkillFed](https://skillfed.io/packages/gigachat) · [View on PyPI](https://pypi.org/project/gigachat/)