Introducing ChatGPT-AccessToken-Web
The ChatGPT-AccessToken-Web project revolutionizes access to the web version of ChatGPT by leveraging an Access Token method. It provides a front end without requiring OpenAI's API quotas. This project is a modification of the ChatGPT-Next-Web project and connects to the GPT-3.5 model by default in the main branch, while there is a separate branch for the GPT-4 model. The necessary backend for this project is the pandoranext project. This endeavor stands on the shoulders of the original developers of ChatGPT-Next-Web and pandoranext. This overview introduces the project, its features, deployment methods, and access considerations.
Key Features
- No OpenAI API Quotas Needed: Remove the stress of consuming limited API call quotas and frequently changing API keys, thus lowering usage costs.
- Comprehensive Markdown Support: Supports LaTeX formulas, Mermaid flowcharts, and code highlighting for enhanced documentation and presentations.
- Elegant User Interface: Features a well-designed responsive UI, dark mode, and PWA support, ensuring a pleasant user experience.
- Fast Loading Speed: Provides exceptionally quick initial load times (~100kb), with streaming response capabilities.
- Privacy Focused: All data is stored locally in the user's browser.
- Predefined Character Roles (Masks): Offers easy creation, sharing, and customization of personalized chat dialogues.
- Extensive Built-in Prompt Library: Provides a wide array of prompts in various languages.
Next Steps
There are currently no specific future plans noted for this project.
Access Token Usage
To use this project, an Access Token is needed. It can be obtained by logging into ChatGPT officially and accessing the session API. An alternative method is also provided to acquire an Access Token.
GPT Plus
Though there is a GPT-4 branch available, no further updates are planned for it.
Account and Password
Only official ChatGPT accounts are supported, excluding third-party logins from platforms such as Google, Microsoft, or Apple. There's also a verification method available without the need for a VPN.
Deployment Instructions
There are several deployment options outlined below:
Deployment Option 1: Using Account and Password
- Obtain a ChatGPT official account and ensure Docker is installed and running.
- Acquire a License ID (accessible online) and configure it in the provided deployment package.
- Follow the Docker commands to set up the backend and web interface.
- Access the web interface through your server's IP at port 3000 in a web browser.
Alternative Deployment 1: Using Access Token
Obtain a JWT URL and configure the deployment with your Access Token. Run the necessary Docker commands to set up access and utilize the provided port for web interaction.
Alternative Deployment 2: Multi-System Architecture Support Using Access Token
Similar to the above options, but with the capability to support various system architectures through Docker configurations. Ensure that the License ID and Access Token are correctly set.
Unsupported Deployment Methods
Vercel and Kubernetes (K8s) deployment methods are not supported.
Open Source License
The project adheres to the Anti 996 License, promoting better working conditions.
Acknowledgments and Additional Notes
This project owes its existence to the efforts of others in the community. Users are encouraged to star the project on GitHub if they find it useful. It humorously asserts that PHP remains the best programming language in the world.