agent-memory

14 tools — 7 read, 7 write.

Tool		What it does
memory_archive	write	Retire one page of your memory that is no longer relevant.
memory_brief	read	Your memory, kept in Mente: your owner's instructions for you, what you're currently tracking, and your pages on the people, clients, projects and lessons you've worked with.
memory_owner_archive	write	The owner archives one page of their own agent's memory, as a new version.
memory_owner_item	read	The owner's view of one memory item: its current text and the metadata of every version.
memory_owner_rail	read	The owner's view of one agent's memory: Instructions, Core memory and the page index.
memory_owner_restore	write	The owner restores an earlier version of one memory item of their own agent, as a new version.
memory_owner_save	write	The owner saves one memory item of their own agent as a new version.
memory_owner_search	read	The owner searches one of their own agent's memory: its pages and Core memory.
memory_owner_version	read	The owner's view of one version of one memory item: its full text and header.
memory_read	read	Open one page of your memory by its id, to see everything on it before you rely on it or change it.
memory_search	read	Search your own memory — your pages and Core memory — for what your page index in memory_brief doesn't show.
memory_update_core	write	Rewrite your Core memory: the short text you keep in mind in every conversation — open threads, fresh corrections, what's true right now.
memory_update_instructions	write	Rewrite your Instructions: who you are, how you work and what to remember — your owner's standing orders, loaded at the start of every conversation and in every runtime you run in.
memory_write	write	Create or rewrite one page of your memory: one page per person, client, project or lesson you work with.

Calling these

Over MCP, find one with search_tools, then run it with execute_tool. Over REST, POST /api/capabilities/<tool> with the arguments as the JSON body. Either way, fetch the input schema for the tool you're about to call rather than guessing — get_tool_schema over MCP, or GET /api/capabilities?name=<tool> over REST.

See How the API works for auth, errors and the shared envelope.