我是一个PM，最近尝试用 Claude Code 去写Markdown 格式的需求文档，它的工作完成得还可以，但当文档规模越来越大的时候，会导致 Tokens 超出限制的问题，也会导致文档结构极度混乱。
Connect Lumina Docs to Claude, Cursor or any other MCP client and it stops being a tab you switch to. 我是一个PM，最近尝试用 Claude Code 去写Markdown 格式的需求文档，它的工作完成得还可以，但当文档规模越来越大的时候，会导致 Tokens 超出限制的问题，也会导致文档结构极度混乱。 好在需求文档是一种结构化的文档，在编写的时候往往只需要关注当前正在编辑的部分，无需每次都阅读全文，正是在这个思路之下，我使用 Claude Code 做了这样一个 MCP Server 它能够解决的问题是： - ✅ 解决大文档token超长问题 - ✅. The lumina docs mcp server is what makes that connection.
Setup follows the usual MCP pattern — install or clone the server, register it in your client's configuration file, restart the client. The configuration blocks on this page cover the common clients.
The toolset is worth reading before you wire it up, because it tells you what the integration is really for:
claude_desktop_config.example.json — 通用模板claude_desktop_config.macos.json — macOS 系统 Python 配置claude_desktop_config.conda.json — Conda 环境配置claude_desktop_config.multi.json — 多 MCP 服务器配置示例document_root — 文档根节点section — 章节（如项目概述、需求说明）module — 功能模块（如总览模块、系统监控模块）feature — 具体功能（如系统状态统计功能）business_flow — 业务流程说明data_display_rules — 数据展示规则permission_rules — 权限控制规则architecture_design — 架构设计Configuration is passed through the environment: PYTHONPATH, DOC_MANAGER_DB_PATH, DOC_MANAGER_EXPORT_DIR, DOC_MANAGER_DATA_DIR, LUMINA_DOCS_SERVER_NAME, DOC_MANAGER_DEBUG, DOC_MANAGER_LOG_LEVEL. Treat anything key-shaped as a real credential — scope it to the minimum the server needs, and rotate it if it ever lands in a shared config.
Plenty of developer tooling servers cover similar ground. The differences that matter in practice are scope of access and how much setup stands between you and a working tool call. Lumina Docs's toolset — claude_desktop_config.example.json, claude_desktop_config.macos.json, claude_desktop_config.conda.json and 11 more — is a fair guide to whether it matches your workflow. It is maintained by GaoZYun; worth a glance at recent repository activity before you build anything load-bearing on it.
We check each listing at SyncDev against the project's documentation before it goes live — if something here drifts out of date, it is a bug worth reporting.
| Tool | What it does |
|---|---|
| claude_desktop_config.example.json | 通用模板 |
| claude_desktop_config.macos.json | macOS 系统 Python 配置 |
| claude_desktop_config.conda.json | Conda 环境配置 |
| claude_desktop_config.multi.json | 多 MCP 服务器配置示例 |
| document_root | 文档根节点 |
| section | 章节（如项目概述、需求说明） |
| module | 功能模块（如总览模块、系统监控模块） |
| feature | 具体功能（如系统状态统计功能） |
| business_flow | 业务流程说明 |
| data_display_rules | 数据展示规则 |
| permission_rules | 权限控制规则 |
| architecture_design | 架构设计 |
| DOC_MANAGER_DB_PATH | <data_dir>/database/documents.db |
| DOC_MANAGER_EXPORT_DIR | ~/Desktop |
然后编辑 Claude Desktop 配置文件：
```json
{
"mcpServers": {
"lumina-docs": {
"command": "/usr/bin/python3",
"args": ["-m", "doc_manager"],
"cwd": "/path/to/lumina-docs",
"env": {
"PYTHONPATH": "/path/to/lumina-docs/src",
"DOC_MANAGER_DB_PATH": "/path/to/lumina-docs/database/documents.db",
"DOC_MANAGER_EXPORT_DIR": "/path/to/exports",
"DOC_MANAGER_DATA_DIR": "/path/to/lumina-docs",
"LUMINA_DOCS_SERVER_NAME": "lumina-docs",
"DOC_MANAGER_DEBUG": "false",
"DOC_MANAGER_LOG_LEVEL": "INFO"
}
}
}
}Configuration as documented by the project. Restart the client after saving.
| Variable | Description | Required |
|---|---|---|
| PYTHONPATH | Filesystem location the server is allowed to use. | Optional |
| DOC_MANAGER_DB_PATH | Filesystem location the server is allowed to use. | Optional |
| DOC_MANAGER_EXPORT_DIR | Filesystem location the server is allowed to use. | Optional |
| DOC_MANAGER_DATA_DIR | Filesystem location the server is allowed to use. | Optional |
| LUMINA_DOCS_SERVER_NAME | Configuration value read at startup. | Optional |
| DOC_MANAGER_DEBUG | Configuration value read at startup. | Optional |
| DOC_MANAGER_LOG_LEVEL | Configuration value read at startup. | Optional |
Kill hallucinated APIs — version-accurate, up-to-date library documentation injected straight into context.
Microsoft's official browser automation server — drive a real browser through the accessibility tree, no screenshots needed.
GitHub's official server — repos, issues, pull requests, Actions and code security, straight from your assistant.
Issue tracking at the speed of conversation — Linear's official hosted server with OAuth and zero install.
Local repository surgery — status, diffs, commits, branches and history for any repo on disk.
Timezone sanity for AI — current time anywhere and correct conversions, without the model doing date math.