As soon as you add other documents into your pinecone vector db this setup will no longer find your handbook anymore. There seems to be a bug in the AI Agent module, because you can use the question&answer chain module that will be working. However the QA chain does not have a context memory, making it useless for a chat-system you intended.
Another good video. I wonder if larger organisations would share their credentials? By the way what recording software are you using and how did set up your recording. I asked because I like the simplicity of it.
Thank you very useful. My BIG question is this (video idea!); I am going to run this off a server in my house, what are the cyber security implications? Can someone trick the AI into giving out my google logins or the SSH key or similar?
As soon as you add other documents into your pinecone vector db this setup will no longer find your handbook anymore. There seems to be a bug in the AI Agent module, because you can use the question&answer chain module that will be working. However the QA chain does not have a context memory, making it useless for a chat-system you intended.
nice video!, how can i get 'metada' of the context?, i mean, the file name of the documents for the answer of the model.
Hi great video: How to get this in a local (company) version??
Another good video. I wonder if larger organisations would share their credentials? By the way what recording software are you using and how did set up your recording. I asked because I like the simplicity of it.
Thank you very useful. My BIG question is this (video idea!); I am going to run this off a server in my house, what are the cyber security implications? Can someone trick the AI into giving out my google logins or the SSH key or similar?