Memory Settings

Memory allows LLMs to actively record important information from conversations, especially user preferences, enabling LLMs to better understand users. Memory is long-term, meaning that even after the conversation ends, the memory still exists, and you can view and edit it.

Enable Memory

Just enable the memory feature, and chat with the assistant. The LLM will automatically record the important information from conversations.

TIP

Make sure the model you are using supports Function Calling feature, and you have turned on the Tools ability in the Providers settings.

Modify Memory

Besides LLM actively updating its memory, you can also manually edit the memory. For example, you can add a new memory: "The user prefers to communicate in a cute tone," so that the LLM will know your preferences.

How it works?

It's based on Function Calling and System Prompt, make sure the model you are using supports these features.

Some reasoning models may not support customize the system prompt, so it can't work with memory.