Welcome to LocalLLaMA! Here we discuss running and developing machine learning models at home. Lets explore cutting edge open source neural network technology together.
Get support from the community! Ask questions, share prompts, discuss benchmarks, get hyped at the latest and greatest model releases! Enjoy talking about our awesome hobby.
As ambassadors of the self-hosting machine learning community, we strive to support each other and share our enthusiasm in a positive constructive way.
Rules:
Rule 1 - No harassment or personal character attacks of community members. I.E no namecalling, no generalizing entire groups of people that make up our community, no baseless personal insults.
Rule 2 - No comparing artificial intelligence/machine learning models to cryptocurrency. I.E no comparing the usefulness of models to that of NFTs, no comparing the resource usage required to train a model is anything close to maintaining a blockchain/ mining for crypto, no implying its just a fad/bubble that will leave people with nothing of value when it burst.
Rule 3 - No comparing artificial intelligence/machine learning to simple text prediction algorithms. I.E statements such as "llms are basically just simple text predictions like what your phone keyboard autocorrect uses, and they're still using the same algorithms since <over 10 years ago>.
Rule 4 - No implying that models are devoid of purpose or potential for enriching peoples lives.
Made some 30 of them talking to the app server and all the containers inside Docker.
Now we can ask how they're all doing and asking application-level questions about records, levels, etc., as well as system level questions like how much RAM the db server is using. Makes for a fun demo. Not sure how useful it will be in production.
Can you explain more about your setup?
I've been playing with something similar where I built a shared file system and a messaging system for my agents, but how are you running your AIs?
I use jan-beta GUI, you can use locally any model that supports tool calling like qwen3-30B or jan-nano. You can download and install MCP servers (from, say, mcp.so) that serve different tools for the model to use, like web search, deep research, web scrapping, download or summarize videos, etc there are hundreds of MCP servers for different use cases.
Never heard of this tool but I’ll check it out.
Mostly I’ve just been making my own dockerfiles and spinning up my own mcp instances.
They’re actually not hard to build so I’m trying to build my own for small utilities, that way I don’t get caught up in an is_even style dependency web.
What does an MCP server do?
Basically it’s a layer to let your LLMs plug into tools.
They generally run on your machine (I use docker to sandbox them) and may or may not call out to useful APIs
One example is I just wrote one to connect to my national weather services RSS feed, so my LLMs can get and summarize the weather for me in the morning.
Works well with Gemma 3n
An MCP server can also just be an interface to something useful but simple like a calculator.