🔌 Connecting to Your Workspace

The Docflo SAP Agent exposes a Model Context Protocol (MCP) server that lets AI assistants like Claude and ChatGPT talk directly to your SAP system. Once connected, you can ask natural language questions and trigger SAP actions straight from your AI workspace.

MCP server URL:

https://api.docflo.ai/sap-mcp

---

Claude

Claude Desktop

Claude Desktop reads its MCP server list from a JSON config file on your machine.

Config file location:

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

Open the file (create it if it doesn't exist) and add the Docflo SAP server under mcpServers:

{
  "mcpServers": {
    "docflo-sap": {
      "url": "https://api.docflo.ai/sap-mcp"
    }
  }
}

If you already have other MCP servers configured, simply add the "docflo-sap" entry alongside them.

Restart Claude Desktop after saving the file. The SAP tools will appear automatically in Claude's tool list.

---

Claude Code (CLI)

Run the following command in your terminal to register the server:

claude mcp add --transport http docflo-sap https://api.docflo.ai/sap-mcp

To verify the server was added:

claude mcp list

You should see docflo-sap in the output. The server is now available in all Claude Code sessions in the current project.

💡 Tip: To add the server globally (available in all projects), add the --scope global flag:

claude mcp add --transport http --scope global docflo-sap https://api.docflo.ai/sap-mcp

---

ChatGPT

ChatGPT Desktop (macOS and Windows) supports MCP servers through a dedicated configuration file.

Config file location:

OS	Path
macOS	~/Library/Application Support/ChatGPT/mcp.json
Windows	%APPDATA%\ChatGPT\mcp.json

Open the file (create it if it doesn't exist) and add the following:

{
  "mcpServers": {
    "docflo-sap": {
      "url": "https://api.docflo.ai/sap-mcp",
      "type": "sse"
    }
  }
}

Restart ChatGPT Desktop after saving. The Docflo SAP tools will be available in new conversations.

⚠️ Note: MCP support in ChatGPT Desktop requires the latest version of the app. If you do not see MCP settings or the tools do not appear after restarting, check for updates in the app or consult the OpenAI Help Center.

---

Microsoft Copilot

Microsoft 365 Copilot connects to MCP servers through the agent management controls in the Microsoft 365 admin center, using an OAuth connection registered by a tenant admin. See Microsoft's official documentation for the full setup flow: Manage Copilot agents and connectors.

When registering the Docflo SAP MCP server, use the following values:

Field	Value
Server URL	https://api.docflo.ai/sap-mcp
Transport	Streamable HTTP (not SSE - the server is stateless HTTP only)
Auth type	OAuth 2.0 (Generic)
Authorization URL	https://api.docflo.ai/sap-mcp/authorize
Token URL	https://api.docflo.ai/sap-mcp/oauth/token
Refresh URL	Same as Token URL
Client ID	Any non-empty string - mcp_oauth ignores whatever is sent and injects the real Auth0 secret server-side
Client secret	Any non-empty string - mcp_oauth ignores whatever is sent and injects the real Auth0 secret server-side
Scope	openid profile email offline_access

💡 Note: The Client ID and Client secret fields are required by the Copilot connector setup form, but their values are not used for authentication. Docflo's mcp_oauth layer discards whatever is submitted and substitutes the real Auth0 application credentials on the server side, so any non-empty placeholder will work.

Once the connection is authorized, the Docflo SAP tools will appear as an available agent/connector for users in the tenant, subject to whatever admin scoping is applied in the Microsoft 365 admin center.

---

After Connecting

Once the MCP server is active in your AI workspace, you can interact with your SAP system using natural language. For example:

• "Show me all open purchase orders for vendor XYZ."
• "Post invoice #1234 to company code 1000."
• "What is the current stock level for material MATL-001?"

The agent will use the services and entities you enabled during the Learning phase to answer your questions and take actions on your behalf.