Connect an AI assistant to DropRFID with MCP
Use the Model Context Protocol (MCP) to let your assistant read scan sessions, browse your SKU catalog, and identify scanned items in your organization.
1. Create an organization API key
Sign in as an organization admin, open API Keys, and select Create API Key. Give this assistant its own key and copy the secret when it appears. The key also grants access to the public API. You can revoke it on the same page.
Admin β API KeysIn Bash or Zsh, run the first line, paste the key at the hidden prompt, and press Enter. Then run the second line to make the key available to your client.
read -rs DROP_RFID_API_KEY export DROP_RFID_API_KEY
2. Connect your MCP client
Use a client that supports Streamable HTTP and a custom Authorization header. For Claude Code, add the following server to .mcp.json in your project root. Start Claude Code from the terminal where you exported the key.
{
"mcpServers": {
"droprfid": {
"type": "http",
"url": "https://www.droprfid.com/api/v1/public/mcp",
"headers": {
"Authorization": "Bearer ${DROP_RFID_API_KEY}"
}
}
}
}For staging, use a staging API key and change the URL to https://staging.droprfid.com/api/v1/public/mcp
3. Check the connection
When Claude Code prompts you, approve the droprfid server. Open /mcp and check that it is connected. Ask βWhat were my last three scan sessions?β An organization with no scans will return an empty session list.
list_recent_sessionsβ Shows the latest sessions with their status and scanned tag counts.list_skusβ Browses the catalog one page at a time.lookup_tagβ Finds the SKU for an RFID EPC or barcode.
If you get HTTP 401, check that the key belongs to the selected environment and is still active. HTTP 429 means the key reached its shared public API limit. Retry after the time in X-RateLimit-Reset.
