About MoonlyBox MoonLink
MoonLink is MoonlyBox’s official MCP service for capturing bookmarks, notes, todos, and web content through AI chats, then searching personal knowledge and memory. Its hosted service shares the same personal memory across clients; the repository provides integration specifications and a behavior template.
Key features
- 01Capture bookmarks, notes, and todos
- 02Save notes and web archives
- 03Search personal knowledge and memory
- 04Organize and soft-delete entries
- 05Authenticate with OAuth or PAT
Use cases
- 01Save web bookmarks in a chat
- 02Record notes and todos
- 03Find saved personal information
- 04Share memory across clients
MCP support
It provides a hosted remote MCP service with tools to capture, search, organize, and delete personal knowledge entries. Connect using OAuth or a PAT.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://moonlybox.cn/mcpclaude mcp add --transport http cn-moonlybox-moonlink https://moonlybox.cn/mcp{"mcpServers":{"cn-moonlybox-moonlink":{"type":"http","url":"https://moonlybox.cn/mcp"}}}Open in Cursor{"servers":{"cn-moonlybox-moonlink":{"type":"http","url":"https://moonlybox.cn/mcp"}}}Open in VS Code{"mcpServers":{"cn-moonlybox-moonlink":{"type":"http","url":"https://moonlybox.cn/mcp"}}}{"mcpServers":{"cn-moonlybox-moonlink":{"serverUrl":"https://moonlybox.cn/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://moonlybox.cn/mcpAuth requiredTools
Details
From the project READMEMoonLink MCP · MoonlyBox
MoonLink is the official MCP (Model Context Protocol) server of MoonlyBox — an AI-native personal knowledge workspace (bookmarks · sticky notes · todos · Study knowledge base · AI memory).
月光链是魔力宝盒的官方 MCP 服务器：AI 对话即采集——收藏网页、存便签、建待办、写书房文档、提取网页存档，并检索个人知识与记忆。云端托管，跨 Claude / Cursor / Cherry Studio / Copilot 等客户端共享同一份个人记忆。
Open spec, hosted service · 规范开放，服务闭源：this repository contains the integration spec & behavior template only. The hosted service is proprietary.
Connect / 接入
Endpoint 端点：https://moonlybox.cn/mcp (Streamable HTTP)
Auth 认证（二选一）：
- OAuth — keyless browser authorization, supported by Claude Desktop / Cherry Studio / CodeBuddy
- PAT — header
Authorization: Bearer mk_...(issue at 站内「AI 托管」页签 → 月光链)
Quick start 快速开始：
- Sign up at https://moonlybox.cn (email registration supported / 支持邮箱注册)
- Open 「AI 托管」→ 月光链 → issue a PAT / or use OAuth in your client
- Add to your client:
{
"mcpServers": {
"moonlink": {
"url": "https://moonlybox.cn/mcp",
"headers": { "Authorization": "Bearer mk_..." }
}
}
}
Tools / 工具（24）
See docs/tools.md for the full table. Highlights:
- Capture 采集: add_bookmark · add_sticky · add_todo · save_note · extract_archive · add_memory
- Query 检索: search_bookmarks · search_library · search_topics · search_memory · list_stickies · list_todos · list_goals · list_library_index · raw_get
- Organize 整理: update_bookmark · update_sticky · update_todo · convert_todo_to_sticky
- Delete（soft, 72h restore for bookmarks）: delete_bookmark · delete_sticky · delete_todo · restore_bookmark
Agent behavior template / 行为规范
docs/agentsTemplate.md — the canonical "Agents.md" template telling AI clients when & how to use these tools (silent memory whitelist, query-first, deletion confirmation discipline). Drop it into your client as AGENTS.md / CLAUDE.md.
Docs / 文档
| Doc | Content |
|---|---|
| docs/tools.md | Full tool table with schemas hints |
| docs/agentsTemplate.md | AI behavior template (AGENTS.md source) |
| docs/oauth-setup.md | OAuth 2.1 / DCR client setup |
| docs/pat.md | Personal Access Token (PAT) guide |
License / 许可
- Integration docs & template: CC BY 4.0
- The hosted MoonLink service & MoonlyBox platform: Proprietary (see NOTICE)
Contact / 联系
- Support & review: [email protected]
- Open-source & community: [email protected]
- Privacy policy: https://moonlybox.cn/privacy
Comments
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.