Free-GPT4-WEB-API

Why we included this project

This self-hosted server gives you one plain-text HTTP endpoint that fronts several modern LLM providers, so you can skip juggling multiple API keys. It routes requests to models like GPT-4, DeepSeek, Claude, and Gemini, and falls back to providers such as Bing and DuckDuckGo when one goes down. You can run it as a Docker container or straight from Python, and the optional browser GUI makes picking a model and provider straightforward. It also handles the messy parts of relying on free endpoints: smart timeouts, retries, provider health tracking, and a blacklist for sources that keep failing. Just don't expect a hard SLA, since the free providers it leans on can get overloaded; treat it as a convenient access layer for prototyping, demos, and personal tooling.