- Execute server commands through a REST API that is compatible with Custom ChatGPT actions
- Interface with external APIs and services.
- Local Tunnel integration for easy access to the server running on a local machine, making the API accessible to ChatGPT.
- Auto-generation of API schema with Swagger is in progress. For a detailed to-do list, please refer to todo.md.
- Node.js v18+
To install the project dependencies, run:
npm install
1. First you will need to install Node.js, at time of writting its v20.16.0 https://nodejs.org/en
npm install
npm run start
5. On the first run, the setup process will guide you through configuring the port, determining whether it runs locally or on a server, and setting the domain.
6. The setup will generate a secret key for use in CustomGPT called authKey, don't share it, it will be used later to allow ChatGPT to call your server or computer
7. Finally, create a CustomGPT here https://chatgpt.com/gpts/editor
8. Add prompt to custom gpt from prompt.md
9. Add your URL to the generated OpenAPI spec, similar to this but with your domain: https://appcookbook.wonderwhy-er.com/openapi.json
For more detailed instructions, please refer to the setup video (TODO: Add video).
Contributions to the Server Commander project are welcome. I did not put in work yet to make it easy to contribute but I will if I see interest in that.
Feel free to reach out to me on
LinkedIn https://www.linkedin.com/in/eduardruzga/
Or Discord https://discord.com/channels/wonderwhyer
Or Twitter/X https://x.com/wonderwhy_er
The project is licensed under the MIT License.