Here's how to setup Jan as an AI provider for Perfect Memory.
In this chapter we will install Jan AI.
Go to the Jan AI https://jan.ai/download and download Jan installer. Run the installer and follow the prompts.
Let's download a model and say Hi to our new AI. Open the Jan and navigate to Hub tab on the left. Now pick a model. We recomend to use at least 8B models. Llama3.1 is a good starting point. Download the llama3.1 8B model. and click Use.
Now we can try chatting with the model. Click use and say Hi in the chat. If it does not reply you might need to download CUDA toolkit from https://developer.nvidia.com/cuda-downloads . Install it and try again. In case of trouble follow Jan troubleshooting faq.
In Jan go to API Server tab on the left. Leave all the default parameters and click Start Server. Be mindfull of the model selected on the right. This is the model that will be available.
In Perfect Memory go to Settings. In the AI group select Jan as a provider. Then expand the AI tab and select the same model you had loaded in Jan. That's it. Now go back to Perfect memory Ai Assistant and try how it works.
This setup makes sure that all of your data never leaves you device when using locl models. Making it perfect for sensitive information.