✨✨ Create your own GPT4 Assistant with ease and use it from the command line. ✨✨
My.Project.mp4
- Runs an OpenAI GPT-4V assistant from the command line
- Execute tools from the command line as required
- Auto-generates JSON schema for your tools from zod validators
- Access it globally from the
ai
command in your terminal
Requirements:
git clone git@github.com:pAIrprogio/gpt-assistant-cli-playground.git
bun install
- Add your OpenAI API key to a .env file using
OPENAI_API_KEY=your-key
- Edit the ai.config.yml file as needed
- This file is loaded from where you run the
ai
command so it can be different for each project. - Autocomplete is available for this file in VSCode using ./schemas/ai.config.schema.json
- This file is loaded from where you run the
To install your assistant globally and access it with the ai
command, run bun link
in the project's folder.
Any change made to the project will be reflected in the global command without extra build step.
Follow the steps in docs/add-new-tool.md
Due to issues in bun-vscode, you need to inspect through an external debugger.
To debug the project, run bun debug
in the project's folder.