Creating a basic AI agent for polling inventory. #478
danielrosehill
started this conversation in
Show and tell
Replies: 1 comment 1 reply
-
Isn't this WAY more complicated and slower than a standard search? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Cleaning Up & Exporting A Static Inventory
So my ultimate objective is to get a proper setup that uses the Homebox API in order to be able to look up items dynamically.
But for a simple proof of concept I decided to do it the dumb way and export my inventory statically.
Step one, export your inventory from the Homebox.
This is totally optional, but I thought it would make sense to do a little bit of cleaning of the data so that irrelevant data for retrieving inventory information could be excluded:
Basically, I got rid of the columns that I thought didn't need to be there.
One thing you can also try is rewriting the headers so that they're a bit more obvious for an AI agent to interpret.
Something like this:
For the purpose of scalability testing, here's where I'm at currently in my inventory:
Once I thought that I'd done enough cleaning, I downloaded a CSV.
Creating Agent In LibreChat
The next step is to create an agent configuration in LibreChat or your preferred large-language model frontend.
The agent is basically a model with a system prompt attached and related to the custom knowledge.
These instructions could almost certainly be improved upon but they should get the job done:
Once you upload your knowledge, the CSV is being vectorized, embedded into a RAG pipeline using the built-in vector database that comes with the stack:
When you can see the file name display, the process has been successful and your agent can now access the inventory:
Make sure to click on save to make sure that the knowledge has been added:
Now our chatbot is ready for use and we can ask you questions about our inventory:
You can see that it's checking the files when asked a question:
It was able to identify all the infrared controllers in my tech cabinet and isolate the one that I was looking for, number 3!
You can ask follow-up questions to retrieve details about the thing you're looking for:
And the chatbot will even respond in the format you requested with the data you wanted and provide a link to the catalog entry:
Beta Was this translation helpful? Give feedback.
All reactions