aledipa/Free-GPT4-WEB-API

What it solves

It provides a way to access modern Large Language Models (LLMs) like GPT-4 and DeepSeek-R1 for free without needing official API keys or paid subscriptions. It solves the problem of high API costs and rate limits by wrapping free providers into a single, self-hosted HTTP interface.

How it works

The project acts as a proxy server that leverages the g4f library to route requests to various free LLM providers (such as Bing and DuckDuckGo). It exposes these models through a simple web API that returns plain text responses. It includes reliability features like automatic provider fallback, health monitoring, and smart timeout handling to ensure responses are delivered even if a specific provider is blocked or overloaded.

Who it’s for

It is designed for developers and hobbyists who want to integrate LLM capabilities into their own applications or workflows (including Siri via Apple Shortcuts) without paying for API access.

Highlights

• Zero Cost: No API keys or subscriptions required for access.
• Flexible Deployment: Can be run via Python source or as a Docker container.
• Management GUI: Includes an optional web-based interface for configuring models, providers, and cookies.
• High Availability: Features automatic switching between providers and a blacklist system for problematic sources.
• Customizable: Supports system prompts, message history, and private mode for restricted access.