Persistent Personal Memory
AI that remembers you, wherever you connect
Remember names, preferences, and meaningful conversations. Connect your website and app so people can continue without explaining themselves again.
Connect your serviceAn example of memory in action
Days later, in a connected app
01 / Why it works
Character, with a personal connection
Less repeating yourself
Names and preferences carry into later conversations, so a new chat can still feel familiar.
Personality, in your language
Shape a character’s personality and speaking style, with conversations in supported languages.
Across your website and app
Connect through the widget and API. Use the same user identity scope to carry new memories across channels.
02 / Inside the memory
How does a conversation become a memory?
A DB stores conversations, names, and preferences. A vector DB helps find memories with a similar meaning to the current conversation.
After saving the conversation, useful details such as names and preferences are organized so related memories can be found in future chats.
Read the flow as text
Your message → identify the user → save the conversation in a DB → organize useful facts → save names and preferences in the personal information store (DB) → make memories searchable in the related memory store (vector DB). Guest demos use session context rather than persistent personal facts.
03 / What to expect
Memory has boundaries
- Use a consistent identityMemory continues for the same character and linked user account. Guest demos use the current session.
- Not every word becomes a memoryUseful information is selected and processing can take time. A new chat may not reflect a memory before it is saved; permanent recall of every conversation is not guaranteed.
- AI can misunderstandExtraction and retrieval can be imperfect. Check important details in memories and replies.