IMPORTANT!

1# IMPORTANT!
2 
3*This project is no longer needed*.  Ragie now supports MCP natively as a streamable HTTP server.  See docs here: [https://docs.ragie.ai/docs/mcp-overview]
4 
5If you would like users within your company to be able to access you knowledge base in applications like Claude or ChatGPT, you may need [MCP Bridge](https://www.ragie.ai/mcp-bridge)
6 
7![image](https://github.com/user-attachments/assets/75e80f87-f39e-4f10-8c97-bbc848bbed82)
8 
9 
10# Ragie Model Context Protocol Server
11 
12A Model Context Protocol (MCP) server that provides access to Ragie's knowledge base retrieval capabilities.
13 
14## Description
15 
16This server implements the Model Context Protocol to enable AI models to retrieve information from a Ragie knowledge base. It provides a single tool called "retrieve" that allows querying the knowledge base for relevant information.
17 
18## Prerequisites
19 
20- Node.js >= 18
21- A Ragie API key
22 
23## Installation
24 
25The server requires the following environment variable:
26 
27- `RAGIE_API_KEY` (required): Your Ragie API authentication key
28 
29The server will start and listen on stdio for MCP protocol messages.
30 
31Install and run the server with npx:
32 
33```bash
34RAGIE_API_KEY=your_api_key npx @ragieai/mcp-server
35```
36 
37### Command Line Options
38 
39The server supports the following command line options:
40 
41- `--description, -d <text>`: Override the default tool description with custom text
42- `--partition, -p <id>`: Specify the Ragie partition ID to query
43 
44Examples:
45 
46```bash
47# With custom description
48RAGIE_API_KEY=your_api_key npx @ragieai/mcp-server --description "Search the company knowledge base for information"
49 
50# With partition specified
51RAGIE_API_KEY=your_api_key npx @ragieai/mcp-server --partition your_partition_id
52 
53# Using both options
54RAGIE_API_KEY=your_api_key npx @ragieai/mcp-server --description "Search the company knowledge base" --partition your_partition_id
55```
56 
57## Cursor Configuration
58 
59To use this MCP server with Cursor:
60 
61### Option 1: Create an MCP configuration file
62 
631. Save a file called `mcp.json`
64 
65* **For tools specific to a project**, create a `.cursor/mcp.json` file in your project directory. This allows you to define MCP servers that are only available within that specific project.
66* **For tools that you want to use across all projects**, create a `~/.cursor/mcp.json` file in your home directory. This makes MCP servers available in all your Cursor workspaces.
67 
68Example `mcp.json`:
69```json
70{
71  "mcpServers": {
72    "ragie": {
73      "command": "npx",
74      "args": [
75        "-y",
76        "@ragieai/mcp-server",
77        "--partition",
78        "optional_partition_id"
79      ],
80      "env": {
81        "RAGIE_API_KEY": "your_api_key"
82      }
83    }
84  }
85}
86```
87 
88### Option 2: Use a shell script
89 
901. Save a file called `ragie-mcp.sh` on your system:
91```bash
92#!/usr/bin/env bash
93 
94export RAGIE_API_KEY="your_api_key"
95 
96npx -y @ragieai/mcp-server --partition optional_partition_id
97```
98 
992. Give the file execute permissions: `chmod +x ragie-mcp.sh`
100 
1013. Add the MCP server script by going to **Settings** -> **Cursor Settings** -> **MCP Servers** in the Cursor UI.
102 
103Replace `your_api_key` with your actual Ragie API key and optionally set the partition ID if needed.
104 
105 
106## Claude Desktop Configuration
107 
108To use this MCP server with Claude desktop:
109 
1101. Create the MCP config file `claude_desktop_config.json`:
111 
112* For MacOS: Use `~/Library/Application Support/Claude/claude_desktop_config.json`
113* For Windows: Use `%APPDATA%/Claude/claude_desktop_config.json`
114 
115Example `claude_desktop_config.json`:
116```json
117{
118  "mcpServers": {
119    "ragie": {
120      "command": "npx",
121      "args": [
122        "-y",
123        "@ragieai/mcp-server",
124        "--partition",
125        "optional_partition_id"
126      ],
127      "env": {
128        "RAGIE_API_KEY": "your_api_key"
129      }
130    }
131  }
132}
133```
134 
135Replace `your_api_key` with your actual Ragie API key and optionally set the partition ID if needed.
136 
1372. Restart Claude desktop for the changes to take effect.
138 
139The Ragie retrieval tool will now be available in your Claude desktop conversations.
140 
141## Features
142 
143### Retrieve Tool
144 
145The server provides a `retrieve` tool that can be used to search the knowledge base. It accepts the following parameters:
146 
147- `query` (string): The search query to find relevant information
148- `topK` (number, optional, default: 8): The maximum number of results to return
149- `rerank` (boolean, optional, default: true): Whether to try and find only the most relevant information
150- `recencyBias` (boolean, optional, default: false): Whether to favor results towards more recent information
151 
152The tool returns:
153- An array of content chunks containing matching text from the knowledge base
154 
155## Development
156 
157This project is written in TypeScript and uses the following main dependencies:
158- `@modelcontextprotocol/sdk`: For implementing the MCP server
159- `ragie`: For interacting with the Ragie API
160- `zod`: For runtime type validation
161 
162### Development setup
163 
164Running the server in dev mode:
165 
166```bash
167RAGIE_API_KEY=your_api_key npm run dev -- --partition optional_partition_id
168```
169 
170Building the project:
171 
172```bash
173npm run build
174```
175 
176## License
177 
178MIT License - See LICENSE.txt for details.
179