# Memory Management

OMate has a long-term memory feature that allows AI characters to remember important information from your interactions, providing a more coherent and personalized conversation experience.

# How Memory Works

When the long-term memory feature is enabled, OMate will:

1. Automatically extract important information about the user from conversations
2. Store this information in the memory repository with date markers
3. In subsequent conversations, the AI can reference these memories to provide more personalized responses
4. You can adjust the types of information to be remembered in custom prompts for more precise memory management

# Enabling Long-Term Memory

Enabling the long-term memory feature is very simple:

1. In the chat interface, click the settings button in the top right corner
2. Go to the conversation management page
3. Turn on the "Enable Long-Term Memory" switch
4. Adjust the maximum number of long-term memories as needed

# Memory Management

# Viewing and Managing Memories

You can view and manage stored memories at any time:

1. On the conversation management page, click the "Manage Long-Term Memory" button
2. On the memory management page, you can:
   • View all stored memory entries
   • Edit the content of memory entries by long-pressing them
   • Swipe left or right to edit or delete individual memories
   • Use the clear button in the top right corner to delete all memories

# Memory Quantity Settings

You can control the number of memories stored by the system:

• Use the slider to adjust the maximum number of long-term memories (100-5000)
• When the number of memories exceeds the set value, the system will automatically delete the oldest memories
• We recommend setting an appropriate memory quantity based on conversation frequency and importance

# Memory Extraction Mechanism

OMate uses specialized prompts to extract and use memories:

# Extracting Memories

After a conversation is completed, the system will automatically:

1. Analyze the recent conversation content
2. Identify new information about the user
3. Extract the new information as concise sentences
4. Save it to the memory repository with a time marker

# Memory Usage

When a new conversation starts:

1. The system loads the user's historical memories
2. These memories are provided to the AI as context
3. The AI will naturally reference these memories at appropriate times
4. Avoiding repeated inquiries about known information

# Best Practices

# Optimizing Memory Quality

• Clearly express important information in conversations to help the system extract it correctly
• Regularly check the memory repository to delete inaccurate or outdated memories
• Edit vague or incomplete memory entries to improve memory quality

# Privacy Note