A small contribution to superlocalmemory

A small contribution to free opensource slm, superlocalmemory. Just an image that was not yet “git lfs”d.

What is SuperLocalMemory?

SLM is a tool to use when coding with the help of AI agents.

Working with this now some time and it really adds a memory to the Large Language Model (LLM). A session that usually starts with a completely blank slate. Every session that you start with the Agent, the context is empty. The Agent only knows about the knowledge that is already fixed in the LLM.

But with SLM, this changes. When SLM is added as a tool to use for the Agent, you add a connection to earlier conversations that you have had in the past. And all stored safely on your local machine. Privacy matters.

Start up the agent within the context of SLM and ask.

• “In this repo, what did we last work on? And what was pending?”
• “Use the same coding style for the unit tests as before” -> without having to explain again the mocking framework we use, why and how we use it.

Or after discarding or explaining why we don’t use certain ways of coding, or not using the LLM suggested code and finally being happy with a solution:

• “remember this way of solving next time we visit similar issues”
• etc. etc.

1. LLMs are just tools to generate random code.
2. LLMs + tools extend capabilities.
3. LLMs + tools + memory: your workflow.

Happy coding!

Fix: convert dashboard screenshot to Git LFS pointer