How MemoryPlugin works

Memories#

A memory is a single discrete fact, preference, or piece of context - "I write in a plain, direct voice," "our backend is Postgres on Neon," "this client prefers weekly updates." You capture memories automatically during a conversation or with one click, and you can read, edit, or delete any of them.

Injection and recall#

When you open an AI tool, MemoryPlugin injects the memories relevant to that conversation. This is Smart Memory - it loads only what is relevant so context stays lean and tokens are not wasted.

Buckets#

Buckets group memories by context - work, personal, a specific project or client - so the right details surface and nothing bleeds across contexts.

Chat history#

Beyond saved memories, MemoryPlugin imports your past conversations from supported platforms and keeps them searchable. See chat history.

Next#

Set it up in Quick setup, or compare ways to connect it in integrations.