Which integrations are available for this server?

Uses environment configuration for storing API keys and settings, with support for fallback mode when API connectivity issues occur.

How do I use mcp-server-zep-cloud?

1. Click on "Install Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-server-zep-cloud search my memory graph for information about my coffee preferences" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

MCP Server for Zep Cloud

Give Claude and Codex long-term memory via Zep Cloud — store conversations, search a knowledge graph, and recall context across sessions.

Prerequisites

Get a free Zep Cloud API key at app.getzep.com.

Related MCP server: mcp-server-cloudbrowser

Quick Start

Claude Code (CLI)

claude mcp add zep-cloud \ -e ZEP_API_KEY=your-key \ -- uvx --from git+https://github.com/fshamim/mcp-server-zep-cloud mcp-server-zep-cloud

Codex CLI

codex mcp add zep-cloud \ -e ZEP_API_KEY=your-key \ -- uvx --from git+https://github.com/fshamim/mcp-server-zep-cloud mcp-server-zep-cloud

Claude Desktop / Codex Desktop

Add to your config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

{ "mcpServers": { "zep-cloud": { "command": "uvx", "args": [ "--from", "git+https://github.com/fshamim/mcp-server-zep-cloud", "mcp-server-zep-cloud" ], "env": { "ZEP_API_KEY": "your-key" } } } }

Setup Script (clone + install)

git clone https://github.com/fshamim/mcp-server-zep-cloud.git cd mcp-server-zep-cloud python scripts/install.py

The script installs the package, prompts for your API key, and configures Claude Desktop automatically.

Available Tools

Tool	Description
`zep_store_memory`	Store content in a memory thread (auto-creates thread)
`zep_get_memory`	Retrieve messages from a thread with pagination and role filtering
`zep_search_memory`	Semantic search across the user's knowledge graph
`zep_get_graph_nodes`	List all entities (nodes) in the knowledge graph
`zep_get_graph_edges`	List all relationships (edges/facts) in the knowledge graph
`zep_get_node_details`	Get detailed info about a node including edges and episodes
`zep_get_thread_context`	Cross-thread context retrieval from past conversations

Docker

docker build -t mcp-server-zep-cloud . docker run -e ZEP_API_KEY="your-key" mcp-server-zep-cloud

Development

Project Structure

core/ run_server.py — Entry point, runs asyncio event loop zep_cloud_server.py — MCP server with 7 tools, stdio transport zep_cloud_client.py — API client wrapping zep-cloud SDK v3 config/ requirements.txt — Dependencies (for manual/Smithery installs) .env.example — Template for environment variables scripts/ install.py — Automated setup script tests/ test_server_tools.py — Unit tests (no API key needed) test_v3_compatibility.py — Integration tests (requires ZEP_API_KEY) pyproject.toml — Package metadata and build config

Install for Development

pip install -e .

Run the Server

mcp-server-zep-cloud # via console entry point python core/run_server.py # direct execution

Run Tests

python tests/test_server_tools.py # Unit tests (no API key needed) python tests/test_v3_compatibility.py # Integration tests (requires ZEP_API_KEY)

License

MIT — see LICENSE.

This server cannot be installed

-

security - not tested

A

license - permissive license

-

quality - not tested

How are these scores calculated?

Resources

GitHub Repository

Need Help?

Report Issue

Related Servers

mcp-server-zep-cloud