albyhub-admin-mcp
MCP server for node-admin operations on an Alby Hub Lightning node via its HTTP API: node info, balances, channels, and sub-wallet apps, with write operations gated behind safety knobs. A generic proxy_request tool covers any endpoint the typed wrappers don't expose.
README
albyhub-admin-mcp
Node-admin operations on your Alby Hub, exposed to an LLM agent. MCP server that wraps the Alby Hub HTTP API — read node info, balances, channels, NWC sub-wallets. Where NWC (nwc-mcp) ends (it's wallet-scoped), this server picks up: hub-wide on-chain balance, channel state, sub-wallet inventory.
v0.1 — defensive design. Alby Hub's admin API has evolved across versions and isn't publicly versioned in a stable way. So v0.1 ships one generic proxy tool that hits any path (the escape hatch) plus typed wrappers around a few well-established endpoints. If a typed wrapper's endpoint-guess fails against your Hub version, the proxy lets you discover the right path without re-shipping. v0.2 hardens the typed wrappers once smoke-tested against real Hubs.
The six tools
| Tool | Method | Path | Purpose |
|---|---|---|---|
albyhub_proxy_request |
* | * | Generic HTTP proxy — escape hatch for any endpoint not yet wrapped. |
albyhub_confirm_request |
* | * | Two-step confirm dispatcher for proxy_request. |
albyhub_get_node_info |
GET | /api/info |
Node identity, network, version. |
albyhub_get_balances |
GET | /api/balances |
On-chain + Lightning aggregate balances. |
albyhub_list_apps |
GET | /api/apps |
NWC connections (sub-wallets) provisioned on this hub. |
albyhub_list_channels |
GET | /api/channels |
Lightning channels with status + capacities. |
All typed wrappers are safe GETs. Non-GET behavior runs through albyhub_proxy_request, which is gated by ALBYHUB_READ_ONLY and ALBYHUB_REQUIRE_CONFIRM.
Requirements
- Node 20+
- A running Alby Hub instance (desktop app on
http://localhost:8080, or self-hosted exposed somewhere) - An Alby Hub API token (Settings → Developer / Apps in the Hub UI)
Install
npx -y albyhub-admin-mcp
Configure
cp .env.example .env
# edit .env: set ALBYHUB_URL (default http://localhost:8080) + ALBYHUB_TOKEN
Required
| Var | Purpose |
|---|---|
ALBYHUB_URL |
Base URL of your Hub. Default http://localhost:8080. Use HTTPS if exposed remotely. |
ALBYHUB_TOKEN |
API access token. Full-scope tokens can drain the hub's on-chain balance — scope down or use ALBYHUB_READ_ONLY=true if your audience matters. |
Optional safety knobs
| Var | Default | Purpose |
|---|---|---|
ALBYHUB_READ_ONLY |
false |
Refuse all non-GET requests via proxy_request. Strongly recommended for first-time setup. |
ALBYHUB_REQUIRE_CONFIRM |
false |
Two-step confirm for non-GET requests. |
ALBYHUB_MAX_REQUESTS_PER_MINUTE |
30 |
Rolling 60s rate limit. |
ALBYHUB_LOG_PATH |
./albyhub-admin-mcp.log |
Server log. |
ALBYHUB_AUDIT_PATH |
./albyhub-admin-mcp-audit.log |
Append-only JSON-line audit log. |
What if my Hub returns 404 on the typed wrappers?
The endpoint path probably differs in your Hub version. Use albyhub_proxy_request to probe — try /api/node, /info, /api/v1/info, etc. Once you find the right path, you can pin it in your usage (and ping the maintainer to fix the typed wrapper in v0.2).
agent: albyhub_proxy_request({ method: "GET", path: "/api/v1/info" })
→ { status: 200, body: { ... } } // found it
Safety model
The proxy tool's pipeline:
ALBYHUB_READ_ONLYgate — non-GET requests blocked outright.- Rate limit — rolling 60s window on the
requestsbucket. ALBYHUB_REQUIRE_CONFIRMgate — non-GET requests return a token;albyhub_confirm_requestexecutes.- HTTP request — Bearer auth, 15s timeout (override via
timeout_ms). - Audit log — every attempt (ok / blocked / error) as one structured JSON line.
GET-only convenience wrappers skip steps 1 + 3 (they're safe by construction) but still go through rate limit + audit.
Companion servers
nwc-mcp— wallet ops via NIP-47. Use this for per-sub-wallet spend; usealbyhub-admin-mcpfor hub-wide / node-level operations.nostr-ops-mcp— NOSTR identity + publishing.marketplace-mcp— NIP-15 marketplace storefront.
License
MIT — see LICENSE.
Contact / Issues
Built by LLMOps.Pro.
- NOSTR:
npub1hdg932jvwc3jdvkqywgqv0ue4nn60exrf92asy8mtazt3hjg7d2s2yw0nw— follow, DM, zap. - Lightning Address:
sovereigncitizens@getalby.com— for support zaps and "this was useful" tips. - Bug reports / feature requests: open a GitHub issue (link forthcoming).
- Security issues: please disclose privately via NOSTR DM before opening a public issue.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。