privateGPT MCP Server

@Fujitsu-AI

Repository

About privateGPT MCP Server

MCP Server including Clients and Agents

Config

No standard config provided

This server doesn't expose a parseable MCP config block in its README. See the repository for install instructions.

Repository

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is privateGPT MCP Server?

A suite of gateway servers from Fsas Technologies AI Team that integrates the PGPT API into any IT environment. It provides three specialized servers: one for legacy systems via TCP, one for local AI agents via STDIO, and one for remote/web architectures via HTTP/SSE.

How to use privateGPT MCP Server?

Clone the repository, select the desired server directory, create a pgpt.env.json with your PGPT API credentials, then follow the local INSTALL.md or run the provided setup script (e.g., ./Install-MPC-Server.sh).

Key features of privateGPT MCP Server

Full coverage of all 23 PGPT API functions
Centralized configuration via pgpt.env.json
Enterprise-grade security: RSA encryption, TLS, SSH key authentication
Optimized for Node.js 20+ with modern ESM modules
Stream-Fix and Auto-Session-Recovery for unstable networks (SSE server)

Use cases of privateGPT MCP Server

Integrate legacy industrial systems via high-performance TCP sockets
Connect local AI tools like Claude Desktop or VS Code via ultra-fast STDIO
Serve remote dashboards and MCP Inspector via resilient HTTP/SSE transport

FAQ from privateGPT MCP Server

Do these servers require an external PGPT instance?

Yes. The MCP Server with STDIO Support and the MCP Server with Streamable-HTTP-Support both require a running PGPT Instance Version 1.5. The API Server with TCP Support also integrates the PGPT API.

What runtime environment is needed?

Node.js 20+ is recommended for full use of modern ESM modules and asynchronous runtimes.

Which transport protocols are supported?

Three protocols: TCP (for legacy systems), STDIO (for local integrations), and Streamable-HTTP via SSE (for remote/web architectures).

How is authentication and encryption handled?

Passwords can be encrypted using RSA public-key cryptography (up to 4096-bit key length). Communication can be secured with TLS (minimum 1.2). Authorization tokens are generated upon successful login and are short-lived.

How do I configure the server?

All servers use a central configuration file pgpt.env.json where you place your PGPT API credentials. Refer to the individual server's INSTALL.md for detailed setup steps.

Frequently asked questions