@sapfever/dashboard-mcp-client

Connect any stdio MCP client to your SAPF Dashboard — giving AI assistants live access to your SAP systems through the dashboard's /api/mcp endpoint.

Zero dependencies. Node.js 18+.

How it works

Any stdio MCP client
       │  
       ▼
@sapfever/dashboard-mcp-client   ←  SAPFEVER_URL + SAPFEVER_TOKEN
       │  
       ▼
SAPF Dashboard
       │  
       ▼
sapf-mcp-server → SAP system

The package is a tiny stdio↔HTTP bridge. It reads line-delimited JSON-RPC from stdin, forwards each message to the dashboard's MCP endpoint authenticated by your API token, and writes the response back to stdout. The dashboard handles spawning and pooling the SAP connection.

OAuth clients (Claude.ai, ChatGPT) connect directly to SAPFEVER_URL/api/mcp without this package — the dashboard exposes a standard Streamable HTTP MCP endpoint with full OAuth 2.1 + PKCE support.

Prerequisites

1. A running SAPF Dashboard instance (self-hosted or cloud)
2. An account on the dashboard with SAP credentials configured
3. An API token — generate one from your profile page under API Tokens
4. Node.js 18 or later

Setup

Claude Desktop

Edit claude_desktop_config.json:

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

{
  "mcpServers": {
    "sapfever": {
      "command": "npx",
      "args": ["-y", "@sapfever/dashboard-mcp-client"],
      "env": {
        "SAPFEVER_URL": "https://your-dashboard.example.com",
        "SAPFEVER_TOKEN": "your-api-token-here"
      }
    }
  }
}

Claude Code

claude mcp add sapfever \
  -e SAPFEVER_URL=https://your-dashboard.example.com \
  -e SAPFEVER_TOKEN=your-api-token-here \
  -- npx -y @sapfever/dashboard-mcp-client

Any other stdio MCP client

Pass SAPFEVER_URL and SAPFEVER_TOKEN as environment variables and set the command to npx -y @sapfever/dashboard-mcp-client. Refer to your client's documentation for how to configure MCP servers.

Microsoft Copilot Studio

Copilot Studio supports Streamable HTTP MCP natively — no package needed. Configure a custom MCP action directly:

| Field | Value | |---|---| | Endpoint URL | https://your-dashboard.example.com/api/mcp | | Authentication | HTTP header | | Header name | Authorization | | Header value | Bearer your-api-token-here |

Claude.ai / ChatGPT connectors

These clients support OAuth and connect directly to the dashboard endpoint — no package needed. Add https://your-dashboard.example.com/api/mcp as a custom connector and sign in when prompted.

Environment variables

| Variable | Required | Description | |---|---|---| | SAPFEVER_URL | Yes | Base URL of your SAPF Dashboard, e.g. https://sap.company.com | | SAPFEVER_TOKEN | Yes | API token from your dashboard profile page |

Generating an API token

1. Log in to your SAPF Dashboard
2. Go to Profile → API Tokens
3. Click Generate Token
4. Choose a name (e.g. Claude Desktop – ECC) and the SAP system
5. Copy the token immediately — it is shown only once

Each token is scoped to a single SAP system. For multiple systems, create a separate token per system and add a separate server entry in your client's config.

Troubleshooting

No tools appear

• Verify SAPFEVER_URL is reachable from your machine
• Check the token is not revoked (profile page → API Tokens)
• Ensure you have SAP credentials for the system the token is scoped to

SAPFEVER_URL and SAPFEVER_TOKEN are required error

• The env vars are not being passed — double-check your client's MCP server config

SAP login failed

• Your SAP password may have changed — update it in the dashboard profile under SAP Credentials, then try again

Tool calls time out

• Some SAP queries are slow; the client allows up to 120 seconds per request
• Check the dashboard logs for the underlying error

License

MIT — Copyright (c) 2026 SAPFever