# Andrea9293 MCP [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/andrea9293/mcp-documentation-server  
**GitHub Stars:** 346  
**npm Downloads (last month):** 579  
**Views:** 6  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/mcp-documentation-server

## Description
Local-first document management and semantic search for AI coding agents

## Tools
Capabilities this server exposes over MCP:

- **add_document** — Add a new document to the knowledge base
- **delete_document** — Delete a document from the collection
- **search_documents** — Search for chunks within a specific document using semantic similarity. Always tell the user if result is truncated because of length. for example if you recive a message like this in the response: 'Tool response was too long and was truncated'
- **get_document** — Use this tool only when user explicitly requests it. Retrieve a specific document by ID. Always tell the user if result is truncated because of length. for example if you recive a message like this in the response: 'Tool response was too long and was truncated'
- **list_documents** — List all documents in the knowledge base
- **search_all_documents** — Search for relevant chunks across ALL documents in the knowledge base using semantic similarity (hybrid: full-text + vector). Useful for cross-document search when you don't know which document contains the answer.
- **get_context_window** — Returns a window of parent content sections around a central parent_index. Use parent_index values from search results. Always tell the user if result is truncated because of length.
- **get_ui_url** — Get the URL of the web UI. use this tool when user ask you to access the web interface, when the user ask you to upload a file or when the user ask you the uploads folder path. All these function are available in the web UI.
- **get_uploads_path** — Get the absolute path to the uploads folder where you can manually place .txt and .md files
- **process_uploads** — Process all .txt and .md files in the uploads folder and create embeddings for them
- **list_uploads_files** — List all files in the uploads folder with their details

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

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

## Documentation

## What Andrea9293 MCP MCP server does

Andrea9293 MCP MCP server provides a local knowledge base for documents used by AI coding agents. It can create documents directly from title and content, inspect stored metadata and previews, retrieve full documents, and delete documents together with their chunks, database records, and associated files.

The server also handles files placed in its uploads directory. Supported upload formats listed by the project are `.txt`, `.md`, and `.pdf`. Agents can find the uploads path, inspect available files, and process the directory into searchable content.

## How it works

Andrea9293 MCP MCP server runs as a FastMCP process over stdio and starts an Express web interface on port 3080 by default. The web interface provides document browsing, uploads, deletion, search, and context exploration. The same operations are available through a REST API under `http://127.0.0.1:3080/api/`.

Documents are split into parent and child chunks before local embeddings are generated with Transformers.js. Orama stores document data, vectorized child chunks, and parent chunks in separate local databases. `search_documents` performs semantic search within one document, while `search_all_documents` combines full-text and vector matching across the collection. `get_context_window` returns nearby chunks around a selected chunk index so an agent can gather surrounding context.

## Setup and configuration

Install the package through an MCP client with `npx`:

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

All documented environment variables are optional. `MCP_BASE_DIR` changes the storage location from its default under the user home directory. `MCP_EMBEDDING_MODEL` selects the embedding model, and models are downloaded on first use. Changing models requires documents to be added again because embeddings from different models cannot be mixed.

Set `START_WEB_UI` to `false` to disable the browser interface. `WEB_HOST` and `WEB_PORT` control its bind address and port. Streaming behavior for large files and the embedding cache can also be configured through environment variables.

## Tools and capabilities

Andrea9293 MCP MCP server exposes tools for:

- Adding, listing, retrieving, and deleting documents.
- Listing upload files and processing them into chunks and embeddings.
- Returning the uploads directory and web UI URL.
- Searching within one document or across all documents.
- Retrieving neighboring chunk context.
- Running Gemini-powered document search when `GEMINI_API_KEY` is available.

The optional `search_documents_with_ai` tool uses Google Gemini rather than only local embeddings. Without that key, the local embedding-based search tools remain available.

## Limitations and notes

The server is designed for local-first operation and does not require an external database or cloud API for its standard embedding workflow. Local embedding models may download approximately 80–420 MB on first use. Storage consists of local Orama files, Markdown copies, and the uploads directory under `MCP_BASE_DIR`.

The default web UI binds to `127.0.0.1:3080`; setting `WEB_HOST` to `0.0.0.0` exposes it on all interfaces. Gemini search depends on a valid `GEMINI_API_KEY` and is not part of the offline-only path.

_Full upstream README: https://allmcps.com/mcp/mcp-documentation-server/readme_