Does the management backend have authentication?
No, the management backend does not implement authorization. Do not expose the service on public servers.
@itcook
基本信息
Base on Graphiti, Enhanced Core Capabilities, Broader AI Model Compatibility, and Comprehensive Management UI.
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"graphiti_pro": {
"transport": "http",
"url": "http://localhost:8082/mcp"
}
}
}工具列表暂未提供
当前收录暂时没有可展示的工具列表,不代表服务没有工具。请查阅项目文档了解支持的能力。
点击上方按钮,直接连接服务器端点实时获取工具列表。
Graphiti Mcp Pro is an enhanced memory repository MCP service and management platform built on Graphiti, a framework for temporally‑aware knowledge graphs designed for AI agents in dynamic environments. It adds asynchronous parallel processing, broader AI model compatibility, and a comprehensive visual Web UI for managing the service.
Deploy via Docker Compose (recommended) by cloning the repository, optionally configuring the .env file, then running docker compose up -d; access the management interface at http://localhost:6062. For manual installation, ensure Python 3.10+, uv, Node.js 20+, Neo4j 5.26+, and an AI model service are available, then run uv sync, configure .env, and start the service with uv run main.py -m (management backend enabled) or uv run main.py (MCP only). The frontend can be built separately using pnpm in the manager/frontend directory.
No, the management backend does not implement authorization. Do not expose the service on public servers.
Only the `streamable-http` transport protocol is supported; `stdio` and `sse` support from the original project has been removed.
When the management backend is enabled, parameters in the `.env` file only take effect on the first startup; subsequent configurations are managed through the backend database.
Yes, through integration with the `instructor` library, Graphiti Mcp Pro supports any model that provides an OpenAI‑compatible API, including DeepSeek, Qwen, and models run via Ollama or vLLM.
The project has not been thoroughly tested due to resource constraints, so it is recommended for personal use only.

See and control a real Windows PC you own — from any MCP client, locally or remotely. Observe (UIA + screenshots), click/type/drag/scroll, launch apps, owner Live View. BYOH: your machine, your key.
🚀 The fast, Pythonic way to build MCP servers and clients.

Handl speaks the Model Context Protocol. Your whole billing operation — projects, invoices, clients, reports, approvals — as tools any AI assistant can call, inside your autonomy rules.
AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

Soar Flight Booking MCP is a remote MCP server available at https://mcp.soar.flights.
评论