Environment setup always seems to take longer than it should. Are you using venv and pip, uv, conda, containers, or something else these days? What makes your setup worth using?
What are you using to manage dev environments in 2026?
Answers
For general Python projects, uv is a strong option. It combines environment and dependency management, and speed is a big part of the appeal. It can also be used for production workflows, so you don't necessarily need a completely separate toolchain for deployment.
That said, it is another tool to install. If pip and venv already cover a small project without much friction, switching isn't mandatory.
Reply to this answer
Plain venv and pip are still a reasonable choice. Keep a requirements file for dependencies, run local tests in the virtual environment, and use a container when you need to check behavior closer to deployment. Simple beats adding tools just because they're new.
Reply to this answer
Conda is worth considering when your project needs packages from conda-forge that aren't available through your usual Python package workflow. The packages you need can matter more than which environment manager feels nicest.
Reply to this answer
Pixi is another option, especially for CUDA or AI work. There's no need to force every project into the same setup if a general Python app and a GPU-heavy project have different dependency needs.
Reply to this answer
For projects with more than Python to manage, mise alongside uv is an interesting combination. It covers a different scope from just picking between pip and Poetry, without going all the way to Nix.
Reply to this answer
Add your answer
Answers are reviewed before publication.
c.356~pocketgyro:
The speed and having fewer separate steps sound useful. I'll compare it with my current pip and venv setup.