Context / Disclaimers # When I say “private ChatGPT” I mean an LLM server / model host + a RAG solution with a web interface. The chat interface has everything I’d need - chat history, the ability to do web searches, connect to MCP servers, and index both uploaded documents and crawled websites. So what’s important, you’re not relying on the model’s “knowledge” but give it the ability to look up the information both from the internet and the resources you give it.
A “20 Questions”-type game implemented on the Pimoroni Picosystem - a Raspberry Pi Pico packaged like a tiny game console.
I got some inspiration from the original patent, prototyped the system in an Elixir Livebook and implemented the game itself in C++.
The livebook stores the bulk of its data on an instance of CloudNativePG deployed to my home kubernetes cluster. The cluster was also useful for feeding the game logic with data - I obtained the reference answers by querying a local Ollama instance using ollamex, which is basically an Elixir implementation of a local-model friendly LiteLLM.
