Read agent memory, sessions, and graphs — then write back with approval
Zep holds the memory behind your agents: sessions, threads, messages, and the knowledge graphs built from them. Luumen reads that history directly — session memory, thread context, node edges, task status — so you can see what an agent actually knew. Changes, like adding a fact triple or a new thread, are previewed and sent for approval first. Connects with an API key.
41 tools: 22 read, 19 write. Reads answer instantly. Writes require approval by default. Everything is logged.
What a governed Zep run looks like inside Luumen.
Authorize once with API token. Luumen lists the scopes each action needs before you approve the connection, and credentials never appear in the chat.
Read actions answer immediately. Anything that writes — add fact triple, add session memory, add thread messages, clone graph, and more — is shown as a plan and requires approval by default, including the 5 actions classified as destructive. Administrators configure that per tool, so you decide exactly which actions can ever run unattended.
You decide. Actions are granted per agent, skill, and team, and per environment — production is not staging. Read access can be broad while writes stay narrow.
Every call to Zep — read or write, approved or declined — is recorded with the actor, the input, and the result, and can be linked to the ticket or change record.
Connect in minutes. Every action scoped, approved, and audited from day one.