This project provides a web-based frontend for ChatGPT, leveraging access tokens instead of OpenAI API keys. It's designed for users who want to bypass API rate limits and reduce costs, offering a feature-rich interface with full Markdown support, a responsive UI, and local data storage.

How It Works

The frontend is a fork of ChatGPT-Next-Web, modified to interact with the pandoranext backend service. This architecture allows users to connect to ChatGPT models (GPT-3.5 and GPT-4 via separate branches) using their official ChatGPT account credentials or an access token, eliminating the need for OpenAI API keys and their associated costs and limitations.

Quick Start & Requirements

• Installation: Docker is the primary deployment method.
• Prerequisites: A ChatGPT official account (username/password) or an access token and JWT URL. Docker must be installed and running.
• Setup: Deployment involves pulling Docker images and running containers, with specific configurations for access tokens or account credentials. Links to obtain access tokens and JWT URLs are provided.
• Resources: Requires a server (local or domestic) for deployment; overseas servers are not necessary.

Highlighted Details

• No OpenAI API key required, bypassing rate limits and reducing costs.
• Full Markdown support including LaTeX, Mermaid, and code highlighting.
• Responsive UI with dark mode and PWA support.
• Streamed responses and fast initial load times.
• Local data storage for privacy.
• Pre-built roles (masks) for personalized conversations and prompt lists.

Maintenance & Community

The project acknowledges its reliance on ChatGPT-Next-Web and pandoranext. No specific community links (Discord, Slack) or roadmap are provided in the README.

Licensing & Compatibility

The project uses the "Anti 996 License," which prohibits 996 work schedules. This license may have implications for commercial use or integration into closed-source projects.

Limitations & Caveats

The GPT-4 branch is not actively updated. The project explicitly states it does not support Google, Microsoft, or Apple third-party logins for ChatGPT accounts. Kubernetes and Vercel deployments are not supported. Access tokens may require re-acquisition and container recreation approximately every 10 days.