stlitepack 0.4.1

A package for converting streamlit apps to stlite apps to remove the need for hosting

stlitepack

stlitepack is a Python utility that helps you turn your existing Streamlit apps into stlite apps — lightweight, browser-only Streamlit apps that run entirely in the client without a server.

With stlitepack, you can:

• 📦 Pack your Streamlit app into a stlite-ready format
• 🗂️ Include additional resources like config.toml
• 🚀 Generate GitHub Actions workflows to (almost) auto-deploy your app to GitHub Pages

📦 Installation

pip install stlitepack

🚀 Usage

from stlitepack import pack

# Pack your Streamlit app (e.g., "app.py") into a stlite bundle
pack("app.py")

This will create an index.html file with all of the required stlite additions, which you can then serve as a static file using a hosting platform such as github pages.

⚠️ Heads Up!

Opening index.html directly in your browser may show navigation or media errors if you use session state in multi-page apps or if you are loading in external file. This is normal — it only happens when using file:// URLs.

Fix: Serve the app over HTTP (e.g., python -m http.server) or deploy to GitHub Pages.

Everything will work correctly once served via a web server!

For local testing prior to deployment, try running

python -m http.server 8000

from the root of your local repository.

Then open http://localhost:8000/index.html (or change the path to reflect the final location of your created index.html file - e.g. http://localhost:8000/docs/index.html)

Documentation

Check out the quickstart guide at sammirosser.com/stlitepack/stlitepack_docs/stlitepack_quickstart.

You can also find the full reference for each function at sammirosser.com/stlitepack/reference/.

🔮 Roadmap

• ✅ v0.1.0: Single-page app packing
• ✅ v0.2.0: Helper functions for GitHub Pages auto-deploy (via GitHub Actions workflow generation)
• ✅ v0.3.0: Multi-page app support (for pages/ subfolder method) and simple additional file inclusion
• ✅ v0.4.0: Lots of improvements!
  • Better support for resource bundling (images, CSVs, assets, config files etc.) of local or web-based file, also leading to multi-page app support via st.navigation() method
  • Starting to automate changes of known stlite/streamlit differences (e.g. await asyncio.sleep(1) being required when using st.spinner)
  • Optional automatic spinning up of a preview server for the deployed app
  • Support for material icons
• v0.4.1: Bugfix for asyncio code
• v0.4.2: Better documentation of v0.4 features
• v0.5.0: Further auto-handling of stlite-specific required changes
• v0.6.0: Add support for generating the required package.json for desktop app bundling
• v0.7.0: Add helpers for generating files for additional deployment options e.g. Docker, Caddy, Nginx, Apache
• v0.8.0: TOML or YAML file support as optional alternative to the packing function
• v1.0.0: Full toolkit for packaging, deploying, and managing stlite apps

Examples of Use

eFIT-tool

All credits go to Peter Saiu and collaborators for the original repository!

• stlite Repository: github.com/Bergam0t/eFIT-tool-stlitepack
• Original Repository: github.com/pete4nhs/eFIT-tool
• Packing script: github.com/Bergam0t/eFIT-tool-stlitepack/blob/main/pack_to_stlite.py
• Hosted stlite app: sammirosser.com/eFIT-tool-stlitepack/

Non-elective Flow Simulation

All credits go to Helena Robinson and collaborators for the original repository!

• stlite Repository: github.com/Bergam0t/hr_Non-Elective-Flow-Simulation
• Original Repository: github.com/Countess-of-Chester-Hospital-NHS-FT/Non-Elective-Flow-Simulation
• Packing script: github.com/Bergam0t/hr_Non-Elective-Flow-Simulation/blob/main/app/pack.py
• Hosted stlite app: http://sammirosser.com/hr_Non-Elective-Flow-Simulation/

🤝 Contributing

Contributions, feature requests, and feedback are welcome!

Open an issue or submit a pull request to help improve stlitepack.

📜 License

Apache 2.0 License. See LICENSE for details.

Acknowledgements

• whitphx for creating the amazing stlite framework!

Generative AI Use Disclosure

This package was developed with the assistance of ChatGPT (OpenAI’s GPT-5 model) as a coding and documentation partner. All code and design decisions were reviewed and finalized by a human, and ChatGPT’s output was used as a foundation rather than a final product.