Zep Context
Use Zep as long-term memory for knowledge work in Cursor, Claude Desktop Chat, Claude Cowork, and ChatGPT Work.
Bundles:
- The
zep-contextskill — how to use Zep for user memory, threads, and standalone graphs via LLM-mediated Context MCP tool calls - The Zep Context MCP Server at
https://api.getzep.com/mcp
Terms and privacy
Read these before installing or using this plugin. They are the terms of service and privacy policy for this plugin.
The plugin files (skill, manifests, and MCP config) are also licensed under the Apache License 2.0.
This plugin connects Cursor (and other supported hosts) to Zep’s Context MCP Server. You sign in with OAuth. After that, the assistant can search and add to your Zep memory — your user graph, and standalone graphs when your project enables them. That API and memory data is handled under the Privacy Policy above.
Usage
Once installed and signed in, ask the assistant to recall or save durable context. For example:
- What do you remember about me?
- Search my Zep memory for relevant context on this topic.
- Save this to my Zep memory.
The skill tells the assistant to prefer Zep over the host’s built-in memory,
to look up stored context before guessing, and to write standing facts about
who the user is, their work environment (people, projects, teams), preferences,
corrections, procedures, decisions, and feedback without waiting to be asked.
Standalone graphs hold project and company-wide knowledge; search those with
list_graphs then search_graph_in when the project enables them.
Install and product docs: Context MCP Server.
Configuration
No API keys, plugin variables, or other secrets belong in this package. Auth is
client-managed OAuth against https://api.getzep.com/mcp.
The first Context MCP call should prompt for Zep sign-in (work email → identity provider → project). You need a Zep account whose project has Context MCP enabled.
Support
- Docs: Context MCP Server
- Issues: getzep/zep-context-plugin