Skip to main content
G4F leverages the powerful curl_cffi library to enable bypassing restrictions and streaming responses from various providers. This document explains how curl_cffi works under the hood and how we utilize it in g4f to enhance the functionality.

What is curl_cffi?

curl_cffi is a Python library that provides a convenient interface to the libcurl C library. It allows making HTTP requests with advanced options and fine-grained control over the request and response handling.
  • Supports all HTTP methods (GET, POST, PUT, DELETE, etc.)
  • Handles cookies and authentication
  • Allows setting custom headers and request bodies
  • Provides proxy support
  • Enables streaming responses

How curl_cffi Works

Under the hood, curl_cffi uses the libcurl C library to perform the actual HTTP requests. It provides a Pythonic interface to interact with libcurl, abstracting away the low-level details.
1

Step 1: Initialize Session

Create a Session object from curl_cffi.requests to manage the HTTP session.
2

Step 2: Configure Request

Set the desired options for the request, such as URL, method, headers, cookies, etc.
3

Step 3: Send Request

Call the appropriate method (get(), post(), etc.) on the Session object to send the request.
4

Step 4: Handle Response

Process the response returned by the server, which can be accessed through the Response object.

Bypassing with curl_cffi in G4F

G4F takes advantage of curl_cffi’s capabilities to bypass certain restrictions imposed by providers. Here’s how it works:
Some providers use Cloudflare protection to prevent automated requests. G4F uses curl_cffi along with a headless browser (e.g., Selenium) to solve the Cloudflare challenge and obtain the necessary cookies and headers to make successful requests.
from g4f.requests import get_session_from_browser

session = get_session_from_browser(url, webdriver, proxy)
To mimic human-like behavior and avoid detection, G4F can impersonate popular browsers like Chrome or Firefox. curl_cffi allows setting custom user agent strings and headers to disguise the requests as originating from a browser.
session = Session(
  headers=headers,
  impersonate="chrome110"
)

Streaming Responses with curl_cffi in G4F

G4F leverages curl_cffi’s streaming capabilities to handle responses that are delivered in chunks. This is particularly useful for providers that support streaming, as it allows processing the response data incrementally without waiting for the entire response to be received.
class StreamResponse:
    def __init__(self, inner: Response):
        self.inner = inner

    async def iter_lines(self):
        async for line in self.inner.aiter_lines():
            yield line
The StreamResponse and StreamSession classes in G4F wrap the curl_cffi response and session objects, respectively, to provide a convenient interface for handling streaming responses.

Conclusion

By leveraging the power of curl_cffi, G4F is able to bypass restrictions, impersonate browsers, and handle streaming responses efficiently. This enables seamless integration with various providers and enhances the overall functionality of the library.
Make sure to install the curl_cffi package to take advantage of these features in G4F:
pip install curl_cffi
I