This is a podcast server application for serving and streaming podcasts in a home lab environment. This system is not designed for a multi million dollar podcast, like Joe Rogan or a podcast hosting site. It's more aimed at Home Labs and Non-profits that want a low cost local hosting of content like a house of worship such as a Masjid or a hobbiest tech enjoyer. Used PCs that can run this setup are relativly cheap -- x86_64 or aarch64. Since it's opensource, you can change the database or use your podcast hosting for your files.
- Podcast subscription, local media
- Streaming capabilities
- Play Audio and Videos via HTML 5
- uv
- wisper (auto installed by dircast)
- ollama (used by dircast)
- Clone the repository:
git clone https://github.com/holz-one/HomeLabPodcastSrv.git
cd PodcastSrv- Install UV
curl -LsSf https://astral.sh/uv/install.sh | sh- Install Ollama (qwen2.5:1.5b)
curl -fsSL https://ollama.com/install.sh | sh
# download model
ollama pull qwen2.5:1.5b
- Import Media:
uv run --refresh dircast.py- Run the application.
uv run app.pyThis project uses a Python development using uv to handel dependancies. I'm using VSCode on a MacBook mid-2012 running Xubuntu 26.04 to develop this system.
I have added the ablity to play HTML5 audio and video content provided in the feed without a propiatary codec like jPlayer. DuckDB is my main DB. I have added some options for Postgres, but it's untested. change the Postgres settings in the .env file.
Mehdi from MotherDuck's DuckDB videos have been very informative; he even helped convince me to switch from DataSpell to VS Code for my IDE via his comments section.
Special thanks also to Manuel Lemos for his websites phpclasses.org and jsclasses.org. He gave me several awards for some of the PHP code I hacked together back in the day, some of which provided the foundational ideas for this project.
I have been using a combination of different AI tools (Leo AI, Gemini, ChatGPT, etc.) to help convert old PHP code into Python and to save time solving complex problems.
This project is licensed under the GPL-3.0 - see the LICENSE.md file for details.