Caute and supermemory.

supermemory is a memory engine that developers build into their agents. Caute is a library a person keeps, which their own assistants read and write. Which one fits depends on whether you are building an agent or using one.

In their words

“supermemory is building the default engine for memory and continual learning for agents.”

From supermemory.ai

Side by side

Caute and supermemory compared
Compared	supermemory	Caute
Who it is for	Developers and teams building agents, through an API, plugins and MCP	A person or team, whose own assistants connect over MCP
Price	Free with $5 of credits a month; Pro $19, Max $100, Scale $399 a month, use metered in credits	Free; Pro $8 a month or $72 a year
How memory changes	Extracted from files, chats, links and data; it updates, merges and forgets on its own	An assistant writes, corrects or forgets a memory; you see each change and can undo a correction

Who it is for: supermemory: Developers and teams building agents, through an API, plugins and MCP; Caute: A person or team, whose own assistants connect over MCP
Price: supermemory: Free with $5 of credits a month; Pro $19, Max $100, Scale $399 a month, use metered in credits; Caute: Free; Pro $8 a month or $72 a year
How memory changes: supermemory: Extracted from files, chats, links and data; it updates, merges and forgets on its own; Caute: An assistant writes, corrects or forgets a memory; you see each change and can undo a correction

Checked 2026-10-25.

What is the same

Both give Claude and other assistants memory over MCP.
Both keep files, links and notes for an assistant to draw on.

Stay with supermemory if

You are building an app or agent and want a memory API for its users.
You want memory extracted from everything and kept up to date without you.

Move to Caute if

You want memory for yourself, read by your own assistants, rather than an API to build on.
You want each memory beside the page or passage it was learned from, with the page kept as it was caught.
You want to see which assistant read which card, and when.
You want a team's spaces to decide what each assistant may read, and spaces kept away from AI entirely.

Moving from supermemory

Caute has no importer for supermemory.

Questions

Does Caute remember things on its own?

No. An assistant writes a memory when it chooses to, and you see it on the Memory page under that assistant's name. A fact it proposes about a card waits for you to keep or dismiss.

Is there an API?

Yes, the same one Caute's own apps use, for reading and writing your own library.

Caute compares itself to products people ask about. Prices and features are theirs, checked on the date shown, and change without notice. Tell us at support@axionarylabs.com if something here is out of date.

Start free