mcpField
S
SurePetcare

SurePetcare

Local
github.com/dirkjanfaberOther0Updated 4w ago

About SurePetcare

SurePetcare is an unofficial MCP server for SurePetcare devices, for users who monitor pets and manage compatible flaps and hubs through the cloud API. It combines pet activity data with device controls.

Key features

  1. 01View pets’ current inside or outside status
  2. 02Inspect pet data and device states
  3. 03Set SureFlap lock states
  4. 04Rename devices and set hub LED mode
  5. 05Review pet activity over a date range

Use cases

  1. 01Check whether a pet is inside or outside
  2. 02Change a cat flap’s lock state
  3. 03Record a pet’s location manually
  4. 04Review inside and outside activity

MCP support

The server exposes pet location, activity reports, device state, SureFlap lock, device renaming, and hub LED controls as MCP tools. It supports stdio use and HTTP mode for remote connectors.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"dirkjanfaber-mcp-server-surepetcare": {
"command": "npx",
"args": [
"-y",
"mcp-server-surepetcare"
],
"env": {
"SUREPETCARE_EMAIL": "<SUREPETCARE_EMAIL>",
"SUREPETCARE_PASSWORD": "<SUREPETCARE_PASSWORD>",
"SUREPETCARE_DEVICE_ID": "<SUREPETCARE_DEVICE_ID>"
}
}
}
}
Run commandnpx -y mcp-server-surepetcare
Environment variables to set
SUREPETCARE_EMAILEmail address of your SurePetcare account
SUREPETCARE_PASSWORDSecretPassword of your SurePetcare account
SUREPETCARE_DEVICE_IDStable UUID identifying this client; a random one is generated on each start if omitted

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmmcp-server-surepetcare0.3.0

Tools

8

Saved tool list · updated 2w ago

This tool takes no parameters.

Details

From the project README

mcp-server-surepetcare

CI

MCP (Model Context Protocol) server for the SurePetcare cloud API. Exposes pet location monitoring, SureFlap lock control, device renaming, and hub LED control as MCP tools.

Disclaimer: This project is not affiliated with, endorsed by, or in any way associated with Sure Petcare Ltd. It is an independent, community-developed integration created by happy users of their hardware and software. SurePetcare, SureFlap, and SureFeed are trademarks of Sure Petcare Ltd. Use of this package is at your own risk. The underlying API is unofficial and reverse-engineered by the community - it may change or break without notice.

Tools

ToolDescription
list_petsList all pets and their current locations (inside/outside)
get_pet_detailsGet raw pet data including microchip tag information
list_devicesList all SurePetcare devices, including live lock state and curfew schedule
set_lock_stateSet the lock state of a SureFlap cat flap
rename_deviceRename a device, re-asserting any explicit lock override so the rename can't silently unlock it
set_pet_locationManually mark a pet inside/outside (e.g. after letting them through a door other than the flap)
set_led_modeSet the hub's LED ring brightness (off/bright/dimmed)
get_pet_reportGet aggregated inside/outside activity stats for a pet over a date range
Lock state values
ValueMeaning
0Unlocked (both directions)
1Locked in (entry only)
2Locked out (exit only)
3Locked (both directions)
LED mode values
ValueMeaning
0Off
1Bright
4Dimmed

Configuration

Set the following environment variables before starting the server:

export SUREPETCARE_EMAIL="[email protected]"
export SUREPETCARE_PASSWORD="yourpassword"
export SUREPETCARE_DEVICE_ID="stable-uuid"   # optional, auto-generated if omitted

Usage with Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "surepetcare": {
      "command": "npx",
      "args": ["mcp-server-surepetcare"],
      "env": {
        "SUREPETCARE_EMAIL": "[email protected]",
        "SUREPETCARE_PASSWORD": "yourpassword"
      }
    }
  }
}

Remote use: claude.ai connector (HTTP + OAuth)

To use the server from claude.ai and the Claude mobile apps, run it in HTTP mode on an always-on machine and add it as a custom connector. claude.ai connects from the internet, so the server needs a public https:// URL. A Cloudflare Tunnel or Tailscale Funnel gives you one without opening ports on your router. Private overlay networks such as ZeroTier or plain Tailscale aren't enough: claude.ai connects from Anthropic's servers, not from your devices.

HTTP mode is single-user: when you connect Claude, the server shows a page asking for your owner passphrase. Anyone who knows it can authorize a client, and anyone with a token can unlock your cat flap - pick a strong one.

npx mcp-server-surepetcare --http   # or MCP_TRANSPORT=http
VariableDefault
MCP_PUBLIC_URL(required)The https:// URL claude.ai reaches the server on
MCP_OWNER_PASSWORD(required)Passphrase for authorizing Claude, at least 12 characters
MCP_HOST127.0.0.1Interface to listen on; the Docker image sets 0.0.0.0
MCP_PORT3000
MCP_STATE_FILE~/.mcp-server-surepetcare/oauth-state.jsonRegistered clients and refresh-token hashes, so Claude stays connected across restarts

plus the SUREPETCARE_* variables from Configuration. Set SUREPETCARE_DEVICE_ID to a fixed UUID, or every restart looks like a new device logging in.

Docker (e.g. on a Raspberry Pi)

Each release publishes ghcr.io/dirkjanfaber/mcp-server-surepetcare for amd64, arm64 and arm/v7 (32-bit Raspberry Pi OS), tagged with the version and latest. Publish its port on loopback only, so the tunnel is the only way in:

# docker-compose.yml
services:
  surepetcare-mcp:
    image: ghcr.io/dirkjanfaber/mcp-server-surepetcare:latest   # or pin a version
    restart: unless-stopped
    env_file: .env   # SUREPETCARE_*, MCP_PUBLIC_URL, MCP_OWNER_PASSWORD
    ports:
      - "127.0.0.1:3200:3000"
    volumes:
      - surepetcare-data:/data
volumes:
  surepetcare-data:

With Tailscale Funnel (no domain needed), run sudo tailscale funnel --bg 3200 and use the URL tailscale funnel status shows as MCP_PUBLIC_URL.

claude.ai only connects on port 443, so Funnel's other ports (8443, 10000) don't work for a connector. If the machine's own name is already taken by another server, give this one its own tailnet machine with a Tailscale container next to it. Drop the ports: mapping above, set MCP_PUBLIC_URL to https://surepet.<tailnet>.ts.net, and add:

  surepetcare-tailscale:
    image: tailscale/tailscale:latest
    hostname: surepet
    restart: unless-stopped
    environment:
      TS_HOSTNAME: surepet
      TS_AUTHKEY: ${TS_AUTHKEY}   # only used for the first login
      TS_STATE_DIR: /var/lib/tailscale
      TS_SERVE_CONFIG: /config/serve.json
      TS_USERSPACE: "true"
    volumes:
      - surepetcare-tailscale:/var/lib/tailscale
      - ./surepet-ts:/config:ro

with surepetcare-tailscale: added under volumes:, and surepet-ts/serve.json:

{
  "TCP": { "443": { "HTTPS": true } },
  "Web": {
    "${TS_CERT_DOMAIN}:443": {
      "Handlers": { "/": { "Proxy": "http://surepetcare-mcp:3000" } }
    }
  },
  "AllowFunnel": { "${TS_CERT_DOMAIN}:443": true }
}

Generate the auth key in the Tailscale admin console under Settings → Keys. Without one, the container prints a login link that expires after about a minute. Once logged in, the machine stays logged in through the volume and the key can be removed.

With a Cloudflare Tunnel (needs a domain on Cloudflare), point a public hostname at the server instead. Don't put Cloudflare Access in front of it: claude.ai can't get past its login page.

Check the public URL before connecting Claude:

curl https://<your public host>/.well-known/oauth-authorization-server

It should return JSON whose issuer matches MCP_PUBLIC_URL.

Update with docker compose pull && docker compose up -d. The server trusts one proxy hop (X-Forwarded-For from the tunnel in front of it) for rate limiting. Don't also publish its port directly to the internet.

Connecting Claude

In claude.ai: Settings → Connectors → Add custom connector, with URL https://<your public host>/mcp. Claude registers itself, opens the passphrase page, and once you allow it, the tools show up in claude.ai and the mobile apps.

References

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.