Run Python Code from Your Git Repository + New Python Versions and Isolated Environments
You can now run your custom Python code directly from your own Git repository and choose from multiple Python versions, including 3.14 beta, with isolated environments to prevent package collisions.
We’ve enhanced the Custom Python Code component with powerful new capabilities to improve flexibility and stability.
Run Code from Your Git Repository
You can now run Python code stored in your own Git repositories, whether public or private. This makes it easier to manage your codebase, version control, and collaborate with your teams using standard Git workflows.
Simply provide the repository URL, and optionally authentication credentials for private repos, and Keboola will fetch and run your code directly—no need to copy-paste code manually into the component. Read the configuration documentation.
What’s New in Runtime Options
You can now select from the following Python runtime environments:
- Python 3.13 – Isolated environment: Latest stable version in a clean, customizable environment. This is the current default.
- Python 3.12 – Isolated environment: Same as above, but with the previous Python version for usage with packages, which still haven't been updated.
- Python 3.14 beta – Isolated environment: Get early access to Python 3.14 features.
- Python 3.10 – Shared environment: Pre-installed packages in legacy versions. This used to be the default choice. It might be a tiny bit faster to start at the cost of running all of the packages in their legacy (3+ years old) versions.
The speed difference when running the legacy 3.10 version with preinstalled Pandas 🐼 compared to installing Pandas everytime in the current 3.13 version was less than 3 seconds in most of our measurements. We really don't think it's worth it – and you're missing out many of the latest features & improvements!
Why it matters
- Isolated environments reduce the risk of package version conflicts by keeping dependencies separate for each run. Although startup takes a few seconds longer, this approach ensures cleaner, more predictable executions.
- The shared environment remains available for backwards compatibility, but it will be slowly rolled out and eventually taken away, so don't hesitate and update your code to keep pace with the development of the Python world.
Choose the best fit for your workloads and ensure greater stability and flexibility when running custom Python code in Keboola.