Skip to main content
For most users, we recommend starting with the public version of Docent. We also provide white-glove hosting support for larger organizations; please reach out if you’re interested.

1. Clone the repo and configure .env

git clone https://github.com/TransluceAI/docent.git
cd docent
cp .env.template .env
You should now have a .env file at the project root. See Environment Variables for details on how to fill it in.
If you’re self-hosting Docent anywhere other than localhost, make sure to set the frontend URL as a CORS origin; e.g., DOCENT_CORS_ORIGINS=http://domain:3001.

2. Start the backend server and frontend UI

Docker Compose is the easiest way to get started, but you may want a manual installation to support faster development loops (e.g., for hot reloading). Finally, try accessing the Docent UI at http://$DOCENT_HOST:$DOCENT_WEB_PORT.

3. Customize the Docent client

When creating Docent client objects, you’ll need to specify custom server and frontend URLs:
import os
from docent import Docent

client = Docent(
    server_url="http://localhost:8889",    # or your own server URL
    frontend_url="http://localhost:3001",  # or your own frontend URL
    api_key=os.getenv("DOCENT_API_KEY"),
)
You’re all set! Check out our quickstart to get started.