Deploy streamlit app that uses pixi

[vmskdl]

I have a streamlit application that uses pixi for environment and package management. In addition to some python packages which can be installed using uv, my app also needs packages that are available only in conda-forge. Using pixi allows me to easily install all dependencies and keep track of them in pixi.toml file automatically. I can launch and run my app locally by running pixi run streamlit run my_app.py without any issues. However, since there is no requirements.txt file, none of my dependencies are installed when I deploy the app to streamlit cloud.

Does streamlit cloud not support pixi in this manner? What are my options?

[AgentStreamy]

Welcome to the Streamlit community and thanks for your thoughtful question! Currently, Streamlit Community Cloud does not support pixi or pixi.toml for dependency management. The platform only recognizes specific dependency files, prioritized as follows: uv.lock, Pipfile, environment.yml (conda), requirements.txt, and pyproject.toml. If none of these are present, your pixi.toml will be ignored and dependencies won’t be installed automatically. See the official App dependencies documentation for details.

Your options are: (1) Export your pixi environment to a supported format like requirements.txt or environment.yml and include it in your repo, or (2) manually create a requirements.txt/environment.yml listing all needed packages. Note that conda-forge-only packages may not always install reliably on Community Cloud due to platform limitations—pip-based installs are preferred. For more, see status and limitations and dependency management. If you hit errors, please share your code/repo and a minimum reproducible example so the community can help debug!

Sources:

• App dependencies - Streamlit Docs
• status and limitations
• dependency management