MCP Documentation Server Server

一個用於文件管理與語意搜尋的伺服器，採用AI嵌入技術並搭配本地JSON儲存。

文件

MCP Documentation Server

Local-first document management and semantic search for AI coding agents. No external databases, no cloud APIs, no vendor lock-in.

Unlike other MCP servers that are CLI-only, this one ships with a full web dashboard — browse, search, upload, and manage your knowledge base from your browser. Every MCP tool is also exposed as a REST API, giving AI agents a lean, schema-free interface.

🏠 Runs fully offline — Orama vector DB with local AI embeddings (Transformers.js)
🌐 Built-in Web UI — starts automatically on port 3080 alongside the MCP server
🔍 Hybrid search — full-text + vector similarity with parent-child chunking
🤖 Optional AI search — Google Gemini for advanced document analysis (bring your own key)
📁 Drag & drop uploads — .txt, .md, .pdf support
📦 Published on the MCP Registry — installable via npx, no clone needed

Quick Start

{
  "mcpServers": {
    "documentation": {
      "command": "npx",
      "args": ["-y", "@andrea9293/mcp-documentation-server"]
    }
  }
}

Open your browser at http://localhost:3080 — the web UI starts automatically.

🤖 Agent Skill (REST API) — recommended for AI agents

Every MCP tool is also accessible via the REST API on http://127.0.0.1:3080/api/. This is the recommended way to interact from AI agents (Claude Code, OpenCode, Gemini CLI, Cursor) because it avoids loading MCP tool schemas into the conversation context — only the response JSON enters.

curl -s http://127.0.0.1:3080/api/config
curl -s http://127.0.0.1:3080/api/documents
curl -s -X POST http://127.0.0.1:3080/api/search-all \
  -H "Content-Type: application/json" \
  -d '{"query": "your search", "limit": 5}'

A ready-to-use skill is included at skills/documentation-server/SKILL.md — it teaches your agent every endpoint with examples. Install it:

npx skills add https://github.com/andrea9293/mcp-documentation-server --skill documentation-server

Basic workflow

Add documents using add_document or place .txt / .md / .pdf files in the uploads folder and call process_uploads.
Search across everything with search_all_documents, or within a single document with search_documents.
Use get_context_window to fetch neighboring chunks and give the LLM broader context.

Web UI

The web interface starts automatically on port 3080 when the MCP server launches. From the web UI you can:

📊 Dashboard — overview of all documents and stats
📄 Documents — browse, view, and delete documents
➕ Add Document — create documents with title, content, and metadata
🔍 Search All — semantic search across all documents
🎯 Search in Doc — search within a specific document
🤖 AI Search — Gemini-powered analysis (if GEMINI_API_KEY is set)
📁 Upload Files — drag & drop files and process them into the knowledge base
🪟 Context Window — explore chunks around a specific index

Configure an MCP client

Minimal

{
  "mcpServers": {
    "documentation": {
      "command": "npx",
      "args": ["-y", "@andrea9293/mcp-documentation-server"]
    }
  }
}

With environment variables (all optional)

{
  "mcpServers": {
    "documentation": {
      "command": "npx",
      "args": ["-y", "@andrea9293/mcp-documentation-server"],
      "env": {
        "MCP_BASE_DIR": "/path/to/workspace",
        "GEMINI_API_KEY": "your-api-key-here",
        "MCP_EMBEDDING_MODEL": "Xenova/all-MiniLM-L6-v2",
        "START_WEB_UI": "true",
        "WEB_HOST": "127.0.0.1",
        "WEB_PORT": "3080"
      }
    }
  }
}

All environment variables are optional. Without GEMINI_API_KEY, only the local embedding-based search tools are available.

MCP Tools

The server registers the following tools (all validated with Zod schemas):

📄 Document Management

Tool	Description
`add_document`	Add a document (title, content, optional metadata)
`list_documents`	List all documents with metadata and content preview
`get_document`	Retrieve the full content of a document by ID
`delete_document`	Remove a document, its chunks, database entries, and associated files

📁 File Processing

Tool	Description
`process_uploads`	Process all files in the uploads folder (chunking + embeddings)
`get_uploads_path`	Returns the absolute path to the uploads folder
`list_uploads_files`	Lists files in the uploads folder with size and format info
`get_ui_url`	Returns the Web UI URL (e.g. http://localhost:3080) — useful to open the dashboard or to locate the uploads folder from the browser

🔍 Search

Tool	Description
`search_documents`	Semantic vector search within a specific document
`search_all_documents`	Hybrid (full-text + vector) cross-document search
`get_context_window`	Returns a window of chunks around a given chunk index
`search_documents_with_ai`	🤖 AI-powered search using Gemini (requires `GEMINI_API_KEY`)

Configuration

Configure via environment variables or a .env file in the project root:

Variable	Default	Description
`MCP_BASE_DIR`	`~/.mcp-documentation-server`	Base directory for data storage
`MCP_EMBEDDING_MODEL`	`Xenova/all-MiniLM-L6-v2`	Embedding model name
`GEMINI_API_KEY`	—	Google Gemini API key (enables `search_documents_with_ai`)
`MCP_CACHE_ENABLED`	`true`	Enable/disable LRU embedding cache
`START_WEB_UI`	`true`	Set to `false` to disable the built-in web interface
`WEB_HOST`	`127.0.0.1`	Bind address for the web UI (use `0.0.0.0` to expose on all interfaces)
`WEB_PORT`	`3080`	Port for the web UI
`MCP_STREAMING_ENABLED`	`true`	Enable streaming reads for large files
`MCP_STREAM_CHUNK_SIZE`	`65536`	Streaming buffer size in bytes (64KB)
`MCP_STREAM_FILE_SIZE_LIMIT`	`10485760`	Threshold to switch to streaming (10MB)

Storage layout

~/.mcp-documentation-server/     # Or custom path via MCP_BASE_DIR
├── data/
│   ├── orama-chunks.msp         # Orama vector DB (child chunks + embeddings)
│   ├── orama-docs.msp           # Orama document DB (full content + metadata)
│   ├── orama-parents.msp        # Orama parent chunks DB (context sections)
│   ├── migration-complete.flag   # Written after legacy JSON migration
│   └── *.md                     # Markdown copies of documents
└── uploads/                     # Drop .txt, .md, .pdf files here

Embedding Models

Set via MCP_EMBEDDING_MODEL:

Model	Dimensions	Notes
`Xenova/all-MiniLM-L6-v2`	384	Default — fast, good quality
`Xenova/paraphrase-multilingual-mpnet-base-v2`	768	Recommended — best quality, multilingual

Models are downloaded on first use (~80–420 MB). The vector dimension is determined automatically from the provider.

⚠️ Important: Changing the embedding model requires re-adding all documents — embeddings from different models are incompatible. The Orama database is recreated automatically when the dimension changes.

Architecture

Server (FastMCP, stdio)
  ├─ Web UI (Express, port 3080)
  │    └─ REST API → DocumentManager
  └─ MCP Tools
       └─ DocumentManager
            ├─ OramaStore          — Orama vector DB (chunks DB + docs DB + parents DB), persistence, migration
            ├─ IntelligentChunker  — Parent-child chunking (code, markdown, text, PDF)
            ├─ EmbeddingProvider   — Local embeddings via @xenova/transformers
            │    └─ EmbeddingCache — LRU in-memory cache
            └─ GeminiSearchService — Optional AI search via Google Gemini

OramaStore manages three Orama instances: one for document metadata/content, one for child chunks with vector embeddings, and one for parent chunks (context sections). All are persisted to binary files on disk and restored on startup.
IntelligentChunker implements the Parent-Child Chunking pattern: documents are first split into large parent chunks that preserve full context (sections, paragraphs), then each parent is further split into small child chunks for precise vector search. At query time, results are deduplicated by parent so that the LLM receives both the matched fragment and the broader context.
EmbeddingProvider lazily loads a Transformers.js model for local inference — no API calls needed.

Development

git clone https://github.com/andrea9293/mcp-documentation-server.git
cd mcp-documentation-server
npm install

npm run dev       # FastMCP dev mode with hot reload
npm run build     # TypeScript compilation
npm run inspect   # FastMCP web UI for interactive tool testing
npm start         # Direct tsx execution (MCP server + web UI)
npm run web       # Run only the web UI (development)
npm run web:build # Run only the web UI (compiled)

Contributing

Fork the repository
Create a feature branch: git checkout -b feature/name
Follow Conventional Commits for messages
Open a pull request

License

MIT — see LICENSE

Support

📖 Documentation
🐛 Report Issues
💬 MCP Community
🤖 Google AI Studio — get a Gemini API key

Star History

Built with FastMCP, Orama, and TypeScript