shruwd-mcp

A local MCP server for Shruwd. It lets an AI agent — Claude Desktop, Claude Code, Cursor, anything that speaks MCP — set up a brand, add the prompts and competitors to track, read how the brand appears in AI-generated answers, and act on the diagnostic findings.

You need an API key from the Shruwd dashboard (Account → API keys). The key is bound to one workspace and reaches your brands in it: every brand if you own the workspace, the brands you are a member of otherwise. It acts with your role on each brand: reads need viewer, writes editor, minting an ingest token admin; creating or archiving a brand needs the owner. To work across several workspaces, mint a key in each and add one server entry per workspace.

Hosted, no install

The same tools run at https://shruwd.io/mcp (Streamable HTTP). In Claude.ai, ChatGPT or Grok, add that URL as a custom connector and sign in with your Shruwd account; no key needed. A client that takes a header can send a key instead:

claude mcp add --transport http shruwd https://shruwd.io/mcp --header "Authorization: Bearer sh_live_…"

Setup for each app: shruwd.io/docs/api/mcp-server.

Claude Desktop

claude_desktop_config.json:

{
  "mcpServers": {
    "shruwd": {
      "command": "npx",
      "args": ["-y", "shruwd-mcp"],
      "env": { "SHRUWD_API_KEY": "sh_live_…" }
    }
  }
}

Claude Code

claude mcp add shruwd -e SHRUWD_API_KEY=sh_live_… -- npx -y shruwd-mcp

Cursor

.cursor/mcp.json:

{
  "mcpServers": {
    "shruwd": {
      "command": "npx",
      "args": ["-y", "shruwd-mcp"],
      "env": { "SHRUWD_API_KEY": "sh_live_…" }
    }
  }
}

Tools

Every tool returns exactly what the API returns. Metrics are never bare numbers: each is a point estimate with its 95% interval and n, or an explicit insufficient_data state, which the tool descriptions tell the agent not to read as zero. Plans that cap visible findings report the remainder as a locked count rather than hiding that it exists.

Environment