GET STARTED
What Persistent Flow is
One memory for every AI assistant you use: saved once, available everywhere, kept in Sweden.
Every new AI chat starts cold. You explain your business, your customers and your decisions again, and none of it accumulates. Persistent Flow is a memory that plugs into the assistants you already use. Save something once in Claude; ask for it later in ChatGPT, Cursor or any other connected client.
It works through the open connector standard (MCP), so there is no new app to learn. You talk to your assistant as usual; it saves to and recalls from your Flow.
What you get
Portable memory
One memory across Claude, ChatGPT, Cursor, VS Code and more. Switch assistants without starting over.
Shared team memory
Workspaces give a team one memory. New people inherit it on day one; people who leave lose access while the knowledge stays.
Provenance
Every answer can say where it came from: the exact words, who wrote them, when, and whether anything has replaced them.
Decisions that are governed
A decision log with who decided and under what mandate, decisions from documents, and meeting minutes signed off by named reviewers.
Communal knowledge
Workspaces where contributions are reviewed before they count — for schools, communities and partner networks.
Sovereign by design
Hosted in Sweden, encrypted per workspace, exportable and erasable. US-operated models only if you opt in.
Three ideas to keep in mind
- Memories are records, not chat logs. Each one has an author, a date and a status. Nothing is silently rewritten; corrections replace, and history keeps both.
- Where you are decides where things go. Your active workspace (or your personal Flow) is where saves land and what search sees.
- Everyday notes are immediate; formal records are reviewed. Saving is instant. Decisions from documents and meeting minutes go through people before they become official.
New here? Go to Connect your assistant, then Your first five minutes.