ResearchifyLabs/py-context-graph ? reverse-engineered prompt

Build me a Python tool that can take messy conversation text like meeting notes, Slack threads, or standup updates and turn it into a clean decision graph.

I want it to find decisions, figure out who made them and what they were about, remove duplicates, add helpful details like topics and key facts, group related decisions together, and let me query the results later. Please make it work with an LLM through a simple adapter, and include a lightweight in memory setup so I can run it locally right away. If it makes sense, also add support for a real backend and a way to plug in a graph database later.

Please include a small example or demo app where I can paste in a conversation, run the pipeline, and see the results in a browser with a timeline, clusters, and a graph view. Use current docs online if you need to, and make sure the whole thing is easy to run from a fresh install.