How is Kumbukum different from Zep?

Zep focuses on agent memory and temporal graph context for AI applications. Kumbukum focuses on shared team knowledge with a web app, notes, memories, URLs, email workflows, Git sync, graph links, API access, self-hosting, and 44 MCP tools.

When should teams choose Kumbukum over Zep?

Choose Kumbukum when people need to capture, inspect, edit, search, and reuse the same memory that AI tools use across Claude, Cursor, ChatGPT, Codex, Windsurf, Zed, and other MCP clients.

Comparison

Kumbukum vs Zep.

Zep is agent memory infrastructure. Kumbukum is a shared knowledge system for teams that need visible, editable context across AI tools.

TLDR:

Should teams use Kumbukum or Zep?

Kumbukum is the better fit when memory is an operational team asset, not only an agent runtime feature. Zep is useful for agent applications that need memory and graph context. Kumbukum is useful when people need to maintain the same knowledge AI uses: notes, memories, URLs, email threads, Git context, projects, graph links, and decisions. Its 44 MCP tools make that library available across Claude, Cursor, ChatGPT, Codex, Windsurf, Zed, and other MCP clients.

Best for Kumbukum

Teams that need visible, editable AI-ready knowledge.
Email, URLs, Git, notes, memories, and graph workflows.
Cross-client MCP access and self-hosting.

Not for Kumbukum

Pure agent app memory with no team workspace.
Developer-only graph memory infrastructure.
Workflows where people never inspect or govern memory.

Need	Kumbukum	Zep
Primary use	Shared team knowledge for AI tools	Agent memory infrastructure
Human editing	Web app and structured objects	Application dependent
Email workflows	Native Email Command Center	Application specific
MCP access	44 tools	Integration dependent
Knowledge graph	Inspectable graph links	Agent graph context