Adding memory to Cursor using MCP and a Knowledge Graph database

I teach a live, interactive program that'll help you build production-ready Machine Learning systems from the ground up. Check it out here:

To keep up with my content:

[LucasWilsonShazbot]

I'm not a dev, but am technical and have worked on software most of my career. Took me all day yesterday, but I made this work in a docker container with my existing app. Thank you!!

[lucasmdo]

Cursor just launched their 1.0 version which has the Memories in beta. Seems they caught up pretty quickly. Still, this is a nice alternative and a sample use of Graphiti.

[tonyhhq]

I'm curious, actually, I haven't tried yet, when memories in the GraphDB accumulated, will this make Cursor Agent Confusing? or How the data is retrieved from the GraphDB?

[neurosync_x]

It has issues pulling values from .env files. Needed to manually assign the env variables (API key, Neo4j password, etc) using the shell/VSCode terminal

[siri_voice]

can I use LM studio? Instead of OPENAI or azure keys.

[realhelmi]

That's an interesting video but I think it misses the main problematic part with AI and knowledge graphs: How does it know what to lookup at what point in time? You can use one single text file as storage if your problem is as simple as just asking a question that you know is stored in the memory.

The Problem with context size and knowledge is that AI doesn't know what it doesn't know. Here's an example: Imagine you're telling it to only run your python script through poetry instead of python directly - like `poetry run main.py` instead of `python main.py` - this will work for a while as long as the context is small but then fall out of context again. Imagine it's storing this in the knowledge graph now. How would it know this is stored there? If it comes to a point where running a python script is what it wants to do... it will just do. how will you have it look up that information by itself without running every request through your knowledge graph multiplle times?

So while having a knowledge graph is a great concept, I didn't find an answer to the underlying problem yet - I hope you did, otherwise I can't imagine how this can really be much more help than a few folders with text files.

[SiselyClovis]

I really appreciate your efforts! I have a quick question: I have a SafePal wallet with USDT, and I have the seed phrase. (wonder obey dial dash soon tank spike scout region undo zero such). Could you explain how to move them to Binance?