Memories
Memories are quick, focused knowledge entries — facts, decisions, context, and learnings that you or your AI tools want to recall later.
Memories vs Notes
Notes are for longer, structured writing. Memories are for bite-sized pieces of knowledge: a decision that was made, a preference to remember, a piece of context that matters. Think of memories as things you'd tell a colleague to get them up to speed quickly.
Memories are especially powerful when used through the MCP Server — your AI assistant can store what it learns in one conversation and recall it in the next.
What You Can Do
Create & Edit
Click New Memory to store something. Give it a title and content. Memories are plain text — no rich formatting, just the information that matters.
Tags
Tag your memories to group them by topic. Tags drive automatic connections in the Knowledge Graph, linking memories to notes and URLs that share the same tags. Use the tag autocomplete to reuse existing tags and keep things consistent.
Source
Memories have an optional source field to track where the information came from — a meeting, a document, a conversation, an AI session. This helps you trace back to the origin when reviewing later.
Project Assignment
Like notes and URLs, every memory belongs to a project. Filter by project in the sidebar to focus on what's relevant.
Search
Memories are indexed with both keyword and vector search. The AI Chat and MCP tools search memories alongside notes and URLs, so stored context surfaces automatically when it's relevant.
Links
Connect a memory to related notes, URLs, or other memories. Open a memory, scroll to the Links section, and search for items to connect. Links appear as removable badges and show up in the Knowledge Graph.
Batch Operations
Select multiple memories with checkboxes to move them between projects or trash them in bulk.
How It Works
When you store a memory, it's saved to MongoDB and indexed in Typesense with a vector embedding. This makes it searchable by meaning — not just exact keywords. Changes are broadcast in real-time to keep all connected clients and tools in sync.