kagi-server MCP Server

1# kagi-server MCP Server
2 
3[![smithery badge](https://smithery.ai/badge/kagi-server)](https://smithery.ai/protocol/kagi-server)
4MCP server for Kagi API integration
5 
6This is a TypeScript-based MCP server that integrates the Kagi Search API. It demonstrates core MCP concepts by providing:
7 
8- Tools for performing web searches and other operations using Kagi's API (currently in private beta)
9 
10## Features
11 
12### Implemented Tools
13- `kagi_search` - Perform web searches using Kagi
14  - Takes a query string and optional limit as parameters
15  - Returns search results from Kagi's API
16 
17### Planned Tools (Not Yet Implemented)
18- `kagi_summarize` - Generate summaries of web pages or text
19- `kagi_fastgpt` - Get quick responses using Kagi's FastGPT
20- `kagi_enrich` - Fetch enriched news results on specific topics
21 
22## Development
23 
24Install dependencies:
25```bash
26npm install
27```
28 
29Build the server:
30```bash
31npm run build
32```
33 
34For development with auto-rebuild:
35```bash
36npm run watch
37```
38 
39## Environment Setup
40 
41Create a `.env` file in the root directory with your Kagi API key:
42 
43```
44KAGI_API_KEY=your_api_key_here
45```
46 
47Make sure to add `.env` to your `.gitignore` file to keep your API key secure.
48 
49## Installation
50 
51### Installing via Smithery
52 
53To install Kagi Server for Claude Desktop automatically via [Smithery](https://smithery.ai/protocol/kagi-server):
54 
55```bash
56npx @smithery/cli install kagi-server --client claude
57```
58 
59To use with Claude Desktop, add the server config:
60 
61On MacOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
62On Windows: `%APPDATA%/Claude/claude_desktop_config.json`
63 
64```json
65{
66  "mcpServers": {
67    "kagi-server": {
68      "command": "/path/to/kagi-server/build/index.js",
69      "env": {
70        "KAGI_API_KEY": "your_api_key_here"
71      }
72    }
73  }
74}
75```
76 
77### Debugging
78 
79Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the [MCP Inspector](https://github.com/modelcontextprotocol/inspector), which is available as a package script:
80 
81```bash
82npm run inspector
83```
84 
85The Inspector will provide a URL to access debugging tools in your browser.
86 
87## Usage
88 
89Once the server is running and connected to Claude Desktop, you can use it to perform web searches. For example:
90 
911. Ask Claude: "Can you search for information about the latest advancements in quantum computing?"
922. Claude will use the `kagi_search` tool to fetch results from Kagi's API.
933. Claude will then summarize or analyze the search results for you.
94 
95Note: The planned tools (summarize, fastgpt, enrich) are not yet implemented and cannot be used.
96 
97## Contributing
98 
99Contributions are welcome! Please feel free to submit a Pull Request. Some areas for contribution include:
100 
101- Implementing the planned tools (summarize, fastgpt, enrich)
102- Improving error handling and input validation
103- Enhancing documentation and usage examples
104 
105## License
106 
107This project is licensed under the MIT License.
108 
109## Roadmap
110 
111- Implement `kagi_summarize` tool for webpage and text summarization
112- Implement `kagi_fastgpt` tool for quick responses
113- Implement `kagi_enrich` tool for fetching enriched news results
114- Improve error handling and add more robust input validation
115- Add more comprehensive usage examples and documentation
116- Publish the package to npm for easy installation and use with Claude Desktop and npx