Installation
Use the Fast Track to understand Docusaurus in 5 minutes ⏱!
Requirements
- uv: A very fast Python package installer and resolver.
- UV is used to install the Dingent CLI and manage Python dependencies.
- Node.js version 18.0 or above (which can be checked by running node -v). You can use nvm to manage multiple Node.js versions on a single machine.
- bun: An all-in-one toolkit for JavaScript and TypeScript applications.
- Bun is used to manage frontend dependencies and run the development server.
Scaffold Project
The easyest way to get started is to use the Dingent CLI to scaffold a new project. This will create a fully functional agent project with all the necessary dependencies and configurations.
uvx dingent init basic
This will prompt you for a project name, author, etc., and then automatically create the project directory and install all frontend and backend dependencies. Then you can navigate to the project directory and start the agent.
Project Structure
If you chose the basic
template and named your project my-awesome-agent
, the project structure will look like this:
├── 📄 README.md
├── 🖥️ frontend/
│ ├── src/
│ │ ├── app/
│ │ └── components/
│ └── package.json
├── ⚙️ backend/
│ ├── main.py
│ ├── config.toml
│ └── langgraph.json
└── 🛠️ mcp/
├── main.py
├── config.toml
├── data/
├── custom_tools/
└── schemas/
Overview of Each Part
-
frontend: This is the part you see and interact with in your browser.
-
backend: This is the project's brain, a Python service.
-
assistants:
Running the development server
To run the development server, you need to set the OPENAI_API_KEY
environment variable and then run the dingent run
command.
# On macOS and Linux
export OPENAI_API_KEY=sk-xxxxxxxxxxxxxxxxxxx # Replace with your OpenAI API Key
# On Windows (PowerShell)
$env:OPENAI_API_KEY="sk-xxxxxxxxxxxxxxxxxxx" # Replace with your OpenAI API Key
uvx dingent run # Must be run in the project directory