pfx MCP - Model Context Protocol Forterro Proffix Px5

pfx MCP - Model Context Protocol Forterro Proffix Px5 is an MCP server that connects AI assistants directly to the Forterro Proffix Px5 ERP system. Built for developers, business analysts, and ERP administrators, it enables Large Language Models like Claude, Gemini, and ChatGPT to interact with corporate ERP data through standard Model Context Protocol interfaces. The server translates natural language requests into structured Proffix REST API operations via JSON-RPC 2.0. Users can query customer master records, check inventory balances, generate sales performance summaries, track open invoices, and manage address data without running manual database queries or exporting spreadsheets. Rather than storing session state or sensitive business records remotely, the server uses parameter-based authentication where Proffix credentials and endpoints are securely supplied via HTTP headers with each request. This bridge provides an extensible layer for automating repetitive ERP workflows, discovering available system endpoints, and orchestrating enterprise operations straight from desktop AI clients or custom software agents.

Category: CRM, ERP & E-commerce

Tags: automation, business-management, erp, proffix

Visit pfx MCP - Model Context Protocol Forterro Proffix Px5

How to install and configure pfx MCP - Model Context Protocol Forterro Proffix Px5

To integrate pfx MCP with an MCP client such as Claude Desktop, configure the connection using an HTTP bridge. 1. Obtain a beta API key from the provider and prepare your Proffix Px5 REST API server URL, port, database, and credentials. 2. In your Claude Desktop configuration file (claude_desktop_config.json), add pfx-mcp under mcpServers: json { "mcpServers": { "pfx-mcp": { "command": "node", "args": ["C:\\mcp\\mcp-http-bridge.js", "https://mcp.pfx.ch/api/server"], "env": { "HTTP_AUTHORIZATION": "Bearer DEIN_API_KEY_HIER", "PROFFIX_USERNAME": "dein-user", "PROFFIX_PASSWORD": "dein-passwort", "PROFFIX_URL": "https://dein-proffix.com", "PROFFIX_PORT": "dein-port", "PROFFIX_DATABASE": "deine-db" } } } } 3. Restart your MCP client to initialize the server handshake and load the px_* operational tools.

What you can do with pfx MCP - Model Context Protocol Forterro Proffix Px5

• Query outstanding customer invoices and filter unpaid balances directly through natural language prompts. - Search warehouse inventory levels to identify items falling below predefined stock thresholds. - Generate aggregated sales reports analyzing top-performing customers and revenue metrics over specified periods. - Create or update company address entries and contact details without opening the ERP interface. - Inspect available ERP endpoints dynamically using discovery tools to automate custom business workflows.

Key facts

• https://github.com/pitwch/pfx-mcp-server
• CRM, ERP & E-commerce
• automation, business-management, erp, proffix

Part of MCP Servers

What is pfx MCP - Model Context Protocol Forterro Proffix Px5?

It is an MCP server that exposes the Forterro Proffix Px5 REST API to AI assistants using the Model Context Protocol. It allows AI clients to run queries, manage master data, and execute ERP operations via JSON-RPC 2.0.

Which MCP clients work with pfx MCP?

The server works with any MCP-compliant client. Supported platforms include Claude Desktop, Cursor IDE, MCP Inspector, custom JavaScript or Python MCP clients, OpenAI API integrations, Google Gemini via API or CLI, and the standalone HTML webview chatbot.

How does authentication work in pfx MCP?

Authentication is parameter-based and stateless. A free API key is passed via Authorization or X-API-Key headers, while Proffix credentials and connection details are transmitted through dedicated HTTP headers per request. The server stores no credentials or persistent session data.

What is the new tool naming convention in pfx MCP?

Starting with recent releases, standard operations are exposed exclusively through tools with the px_* prefix. Legacy tools such as proffix_search_endpoints are disabled by default but can be temporarily enabled by passing the X-Legacy-Mode header with a value of 1.