SurePetcare

io.github.dirkjanfaberv0.3.0更新於 Oct 3, 2026

Unofficial MCP server for SurePetcare devices: SureFlap pet/cat flaps and SureFeed feeders.

概覽

AI 產生的概覽

非官方 MCP 伺服器,讓助理讀取 SurePetcare 寵物位置並控制 SureFlap 門鎖、餵食器與集線器 LED。

功能
連接非官方的 SurePetcare 雲端 API,提供寵物與裝置相關工具。可列出寵物及其室內/室外位置,取得包含晶片標籤資訊的原始寵物資料,列出裝置及即時鎖定狀態與宵禁排程,設定 SureFlap 貓門鎖定狀態,重新命名裝置,手動標記寵物在室內或室外,設定集線器 LED 環亮度,並依日期範圍回傳室內/室外活動統計。
適用情境
適合擁有 SureFlap 貓門、SureFeed 餵食器或 SurePetcare 集線器的使用者,希望助理查看寵物位置、調整鎖定狀態或取得活動報告。這是社群整合,並非 Sure Petcare 官方產品。
執行需求
透過 npm 套件 mcp-server-surepetcare 以 stdio 在本機執行(需要 Node.js/npx),也可用 HTTP 模式接入 claude.ai 連接器。需要 SurePetcare 帳號電子郵件 SUREPETCARE_EMAIL 與密碼 SUREPETCARE_PASSWORD;SUREPETCARE_DEVICE_ID 為選用。HTTP 模式另需 MCP_PUBLIC_URL 與 MCP_OWNER_PASSWORD,以及公開 HTTPS 網址和網路存取。
安裝前請注意
需要把 SurePetcare 帳號密碼放在 SUREPETCARE_PASSWORD 中,請妥善保管。工具會改動真實硬體:set_lock_state 可上鎖或解鎖貓門,rename_device 與 set_led_mode 會變更裝置設定,set_pet_location 會覆寫記錄的位置。HTTP 模式下,任何知道擁有者通關密語或持有權杖的人都能授權用戶端並解鎖貓門,請使用強度足夠的 MCP_OWNER_PASSWORD,且不要把連接埠直接暴露到網際網路。此 API 為非官方逆向工程,可能隨時失效。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 SurePetcare,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Desktop only,透過 STDIO。 STDIO 服務會啟動本機處理程序,因此需要 SourceWeft 桌面主機。

其他 MCP 客戶端

參照 儲存庫 中的啟動說明。

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:

bash
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:

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.

bash
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:

yaml
# docker-compose.ymlservices:  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:/datavolumes:  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:

yaml
  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:

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:

bash
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

來源:README.md,提交 590e092

工具

0
工具後設資料尚未被收錄。

版本歷史

1
  1. v0.3.0最新Oct 3, 2026