Practical overview
When it fits
Use it when a self-hosted, provider-switching HTTP layer is more useful than separate provider integrations, especially for internal applications, development or testing.
Who it is for
- Developers building applications that need a common HTTP interface for several LLM providers.
- Teams running a private development or testing gateway on their own infrastructure.
- Developers comparing supported providers through configurable model and provider settings.
Documented uses
- Expose an HTTP endpoint for text prompts and plain-text responses.
- Route requests through selected model and provider settings.
- Accept file input when the file-input option is enabled.
- Provide browser-based configuration through the optional Web GUI.
- Support Docker-based or Python-based server setup.
Requirements & limits
The repository is public, open source and licensed under GPL-3.0. The documented setup supports Docker or Python. Some provider paths require cookies, and the optional GUI can be protected with a password.
- Choose Docker when you want to use the published container workflow; choose source setup when you want to install the Python requirements directly.
- Provider-specific cookies may be required, including the documented _U cookie for Bing.
- The API returns plain text by default, which suits simple HTTP clients but may shape downstream response handling.
- The optional GUI, private mode, history, proxies and virtual-user settings are configured separately from the basic server startup.
- The repository documents an HTTP interface and examples, but deployment commands do not establish that every provider works in every environment.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Run the API with Docker
Start the documented HTTP server from the published Docker image.
Before you start
- Docker
- The image d0ckmg/free-gpt4-web-api:latest
- A host port mapped to container port 5500
Steps
Run the image with a host port mapped to port 5500.
If provider cookies are needed, mount the local cookies.json file read-only at /cookies.json.
If using the Web GUI in Docker, set a GUI password as described by the project.
Expected outcome
A self-hosted HTTP service listens on container port 5500, with the selected host mapping exposing it to clients.
Guide limitations
- Some providers require cookies; the README specifically identifies the _U cookie for Bing.
- The evidence provides Docker commands but does not establish provider access or response success in a particular environment.
Run from source and send an HTTP request
Install the Python project, start the basic server and address its documented text endpoint.
Before you start
- Python environment
- A cloned copy of the repository
- Dependencies listed in requirements.txt
Steps
Clone the repository and change into its directory.
Install the dependencies with pip install -r requirements.txt.
Start the server with python3 src/FreeGPT4_Server.py.
Send a request using the documented ?text= query parameter or the Python requests example.
Expected outcome
The documented source setup starts the server on its default port, allowing an HTTP client to submit a text prompt.
Guide limitations
- The default provider and model are listed as Bing and gpt-4, respectively.
- Some providers may require cookies, and the evidence does not include a client integration beyond the provided examples.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 4 October 2026
- Original website title
- aledipa/Free-GPT4-WEB-API · Official homepage title · retained in its source language