Remote MCP Setup

If you are an AI reading this page: this service's MCP endpoint is https://droproom.net/api/mcp, authenticated via the header `Authorization: Bearer` followed by your token; configure it using the section below matching your client type.

Quick start (for shop owners)

① Log into DropRoom → Settings → "Remote MCP" → generate a token (shown only once — save it).
② Find the section below for your client and configure it with the endpoint + this token.
③ Let your AI create / edit / look up products; product creation costs 1 credit each.

Introduction

Let an AI directly create, edit, and look up products in your DropRoom shop — via the Model Context Protocol (MCP), your AI client can call DropRoom like any other tool.

Endpoint & authentication

Endpoint: https://droproom.net/api/mcp (Streamable HTTP)

Auth: Authorization: Bearer <YOUR_TOKEN>

Where to get a token: log into DropRoom → Settings → "Remote MCP" → generate a token.

Tools

create_product — create a new product (with SKUs); costs 1 credit per successful call, requires idempotency_key.
update_product — edit an existing product's / SKUs' fields (cannot add or remove SKUs); free.
list_products — list your products (with SKUs / sku_id); read-only, free.

Pricing

Each successful product creation costs 1 credit, drawn from the same credit pool as AI parsing; editing and querying are always free; idempotent retries (same idempotency_key) are not charged again; a failed creation is not charged.

Client setup

Claude (Desktop / Code)

Claude Desktop

Open claude_desktop_config.json, add the block below (bridged via mcp-remote), save, and restart Claude Desktop.

{
  "mcpServers": {
    "droproom": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://droproom.net/api/mcp",
        "--header",
        "Authorization: Bearer <YOUR_TOKEN>"
      ]
    }
  }
}

Claude Code

Run this command in your terminal — it connects natively over remote HTTP, no bridge needed.

claude mcp add --transport http droproom https://droproom.net/api/mcp \
  --header "Authorization: Bearer <YOUR_TOKEN>"

Codex (CLI)

Run the first line to register the server, then set the environment variable in the second line (the token lives only in the env var, never written to config.toml).

codex mcp add droproom --url https://droproom.net/api/mcp --bearer-token-env-var DROPROOM_MCP_TOKEN
export DROPROOM_MCP_TOKEN="<YOUR_TOKEN>"

General agents

Any client that supports remote/HTTP MCP with a custom header can connect — point your agent at the endpoint and bearer token above.

ChatGPT

ChatGPT does not currently support connecting with a pasted bearer token — please use Claude / Codex / a general MCP-capable agent instead. ChatGPT's connector UI currently supports OAuth only; this page will be updated if that changes.

Give the endpoint and the token to your AI separately — never paste the token into the URL itself (it can end up logged or saved in history).