Dynamic image building with binderhub

This gives users the ability to build and share their own user environment images from the JupyterHub UI.

When enabled, users can:

1. Provide a link to a GitHub repository

2. Wait for BinderHub to build an image from that repository

3. Launch their server with the freshly built image

This is particularly useful for:

• Research reproducibility: Users can specify exact environments from published repositories

• Educational settings: Instructors can provide course-specific repositories

• Sharing work with colleagues: Users can share the computational environment in addition to notebooks and computational content.

Enable dynamic image building¶

Dynamic image building requires two components:

1. A BinderHub service configured in your JupyterHub

2. The dynamic_image_building flag enabled in your profile options

Example configuration¶

The repository includes working example configurations for local development:

• jupyterhub_config.py - Shows how to configure BinderHub as a JupyterHub service and enable the dynamic_image_building flag in profile options

• binderhub_config.py - Basic BinderHub configuration for local development

When to use this vs. the binderhub ui¶

Use jupyterhub-fancy-profiles with BinderHub integration when you’re building a persistent JupyterHub with:

• Persistent home directories

• Multiple profile options

• Strong access control

• User authentication

Use the standard BinderHub UI when you’re building an ephemeral hub where:

• Users click a link for immediate, temporary access

• No persistent storage is needed

• Sessions are short-lived

• Anonymous or lightweight auth is sufficient