ReplenishRadar MCP Server

Agent-ready inventory intelligence for Shopify and Amazon sellers. Stockout risk, demand forecasts, purchase order recommendations, and real-time alerts — all accessible to AI agents via the Model Context Protocol.

What is it?

The ReplenishRadar MCP server lets AI agents (Claude Desktop, custom agents, n8n workflows) query live multi-channel inventory data, request purchase order recommendations, create and manage POs, and trigger inventory syncs — with full human-in-the-loop safeguards.

All write operations create drafts only. No agent can send a PO to a supplier without explicit human approval.

Quick Start (Claude Desktop)

1. Get an API key from ReplenishRadar > Settings > API Keys
2. Add to your Claude Desktop config (~/.claude/claude_desktop_config.json):

{
  "mcpServers": {
    "replenishradar": {
      "command": "npx",
      "args": ["-y", "@replenishradar/mcp-server"],
      "env": {
        "REPLENISHRADAR_API_KEY": "rr_sk_live_your_key_here"
      }
    }
  }
}

1. Restart Claude Desktop
2. Ask: "What are my top stockout risks?" or "Draft POs for everything critical"

Tools (28)

Read Tools (18) — Growth+ tier

Write Tools (10) — Growth+ creation, Scale operations

Setup Patterns

Claude Desktop (5 minutes)

See Quick Start above.

Custom Agent (Python + Claude API)

import anthropic
from replenishradar import ReplenishRadarMCPClient

rr = ReplenishRadarMCPClient(api_key="rr_sk_live_...")
client = anthropic.Anthropic()
tools = rr.get_tools()

response = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=4096,
    tools=tools,
    messages=[{"role": "user", "content": "Draft POs for my critical SKUs"}],
    system="You are an inventory operations agent. Create draft POs only."
)

OpenClaw + Slack/Discord

1. In ReplenishRadar: Settings > API Keys > Create (scope: write)
2. In OpenClaw: Tools > Add MCP Server > URL: https://api.replenishradar.com/mcp, Auth: Bearer rr_sk_live_...
3. Configure webhook: Settings > API & Webhooks > Subscribe to events
4. Agent receives stockout alerts and can act on them in Slack/Discord

n8n / Make (No-Code)

Call the REST API directly at https://api.replenishradar.com/api/mcp/call with your API key in the Authorization header.

For detailed setup guides and system prompt templates, see docs/SETUP.md.

Rate Limits

Security

• Org-scoped API keys — cross-org access impossible at the query layer
• Draft-only enforcement — server-side, not client-side
• Circuit breaker — suspends keys exceeding 200% of hourly limit
• Audit logging — every tool call logged with 7-year retention
• Amazon BSA compliant — identification headers, kill switch, no training on Amazon data
• OWASP LLM Top 10 — controls for prompt injection, excessive agency, sensitive data exposure

For full security documentation, see docs/SECURITY.md.

Architecture

Your agent (Claude, GPT, Gemini, custom)
         |
         v
@replenishradar/mcp-server (stdio bridge)
         |
         v
POST https://api.replenishradar.com/api/mcp/call
         |
         v
ReplenishRadar backend (rate limiting, audit, org-scoping)
         |
         v
Your inventory data (Shopify + Amazon)

The MCP server is a thin stdio bridge that dispatches all tool calls to the ReplenishRadar REST API. Authentication is via REPLENISHRADAR_API_KEY environment variable.

Environment Variables

Documentation

• Setup Guide & Agent Patterns
• Security & Compliance
• MCP Setup Blog Post
• API Docs
• Website

License

Proprietary — see LICENSE