Use Channel Mirrors and Air-Gapped Package Servers

In environments without direct internet access, or where conda channels must be served from internal servers, the Python Environment Provider node can be pointed at channel mirrors via Pixi's global configuration files. No changes to workflows or node settings are needed — the configuration is picked up both when resolving dependencies in the node dialog and when the environment is created during execution.

Configure a mirror

Place a config.toml file in one of Pixi's standard locations:

Scope	Linux	macOS	Windows
System-wide	/etc/pixi/config.toml	/etc/pixi/config.toml	C:\ProgramData\pixi\config.toml
Current user	~/.config/pixi/config.toml	~/Library/Application Support/pixi/config.toml	%APPDATA%\pixi\config.toml

For KNIME Executors, which typically run under a service account, the system-wide location is recommended. In a containerized executor, add the file to the image or mount it into the container.

An example configuration that redirects the conda-forge and knime channels to an internal mirror:

[mirrors]
"https://conda.anaconda.org/conda-forge" = [
    "https://my-internal-server.example.com/mirrors/conda-forge"
]
"https://conda.anaconda.org/knime" = [
    "https://my-internal-server.example.com/mirrors/knime"
]

INFO

Lock files always record the canonical channel URLs — the mirror substitution happens when files are downloaded. This means a workflow whose dependencies were resolved on a machine with internet access can be executed in the air-gapped environment and vice versa, without re-resolving.

Further options

The same configuration files support additional options that are useful behind corporate infrastructure:

• authentication_override_file for mirrors that require authentication, e.g. when the system keyring is not available to the executor's service account.
• [repodata_options] for repository servers that do not support all repodata formats (e.g. some Artifactory versions).
• tls_no_verify and proxy-related settings for TLS-intercepting corporate proxies.

See the Pixi configuration reference for the complete list.

Next steps

• Use the Python Environment Provider (Pixi): Set up a portable environment that works on desktop and web.
• Python Environments on Executors: Understand the environment types available on KNIME Executors.
• Troubleshooting: Solutions for network and installation failures.