Cross-LLM MCP Server

1# 🤖 Cross-LLM MCP Server
2 
3> **Access multiple LLM APIs from one place.** Call ChatGPT, Claude, DeepSeek, Gemini, Grok, Kimi, Perplexity, and Mistral with intelligent model selection, preferences, and prompt logging.
4 
5An [MCP (Model Context Protocol)](https://modelcontextprotocol.io) server that provides unified access to multiple Large Language Model APIs for AI coding environments like Cursor and Claude Desktop.
6 
7[![Trust Score](https://archestra.ai/mcp-catalog/api/badge/quality/JamesANZ/cross-llm-mcp)](https://archestra.ai/mcp-catalog/jamesanz__cross-llm-mcp)
8 
9## Why Use Cross-LLM MCP?
10 
11- 🌐 **8 LLM Providers** – ChatGPT, Claude, DeepSeek, Gemini, Grok, Kimi, Perplexity, Mistral
12- 🎯 **Smart Model Selection** – Tag-based preferences (coding, business, reasoning, math, creative, general)
13- 📊 **Prompt Logging** – Track all prompts with history, statistics, and analytics
14- 💰 **Cost Optimization** – Choose flagship or cheaper models based on preference
15- ⚡ **Easy Setup** – One-click install in Cursor or simple manual setup
16- 🔄 **Call All LLMs** – Get responses from all providers simultaneously
17 
18## Quick Start
19 
20Ready to access multiple LLMs? Install in seconds:
21 
22**Install in Cursor (Recommended):**
23 
24[🔗 Install in Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=cross-llm-mcp&config=eyJjcm9zcy1sbG0tbWNwIjp7ImNvbW1hbmQiOiJucHgiLCJhcmdzIjpbIi15IiwiY3Jvc3MtbGxtLW1jcCJdfX0=)
25 
26**Or install manually:**
27 
28```bash
29npm install -g cross-llm-mcp
30# Or from source:
31git clone https://github.com/JamesANZ/cross-llm-mcp.git
32cd cross-llm-mcp && npm install && npm run build
33```
34 
35## Features
36 
37### 🤖 Individual LLM Tools
38 
39- **`call-chatgpt`** – OpenAI's ChatGPT API
40- **`call-claude`** – Anthropic's Claude API
41- **`call-deepseek`** – DeepSeek API
42- **`call-gemini`** – Google's Gemini API
43- **`call-grok`** – xAI's Grok API
44- **`call-kimi`** – Moonshot AI's Kimi API
45- **`call-perplexity`** – Perplexity AI API
46- **`call-mistral`** – Mistral AI API
47 
48### 🔄 Combined Tools
49 
50- **`call-all-llms`** – Call all LLMs with the same prompt
51- **`call-llm`** – Call a specific provider by name
52 
53### ⚙️ Preferences & Model Selection
54 
55- **`get-user-preferences`** – Get current preferences
56- **`set-user-preferences`** – Set default model, cost preference, and tag-based preferences
57- **`get-models-by-tag`** – Find models by tag (coding, business, reasoning, math, creative, general)
58 
59### 📝 Prompt Logging
60 
61- **`get-prompt-history`** – View prompt history with filters
62- **`get-prompt-stats`** – Get statistics about prompt logs
63- **`delete-prompt-entries`** – Delete log entries by criteria
64- **`clear-prompt-history`** – Clear all prompt logs
65 
66## Installation
67 
68### Cursor (One-Click)
69 
70Click the install link above or use:
71 
72```
73cursor://anysphere.cursor-deeplink/mcp/install?name=cross-llm-mcp&config=eyJjcm9zcy1sbG0tbWNwIjp7ImNvbW1hbmQiOiJucHgiLCJhcmdzIjpbIi15IiwiY3Jvc3MtbGxtLW1jcCJdfX0=
74```
75 
76After installation, add your API keys in Cursor settings (see Configuration below).
77 
78### Manual Installation
79 
80**Requirements:** Node.js 18+ and npm
81 
82```bash
83# Clone and build
84git clone https://github.com/JamesANZ/cross-llm-mcp.git
85cd cross-llm-mcp
86npm install
87npm run build
88```
89 
90### Claude Desktop
91 
92Add to `claude_desktop_config.json`:
93 
94**macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`  
95**Windows**: `%APPDATA%\Claude\claude_desktop_config.json`
96 
97```json
98{
99  "mcpServers": {
100    "cross-llm-mcp": {
101      "command": "node",
102      "args": ["/absolute/path/to/cross-llm-mcp/build/index.js"],
103      "env": {
104        "OPENAI_API_KEY": "your_openai_api_key_here",
105        "ANTHROPIC_API_KEY": "your_anthropic_api_key_here",
106        "DEEPSEEK_API_KEY": "your_deepseek_api_key_here",
107        "GEMINI_API_KEY": "your_gemini_api_key_here",
108        "XAI_API_KEY": "your_grok_api_key_here",
109        "KIMI_API_KEY": "your_kimi_api_key_here",
110        "PERPLEXITY_API_KEY": "your_perplexity_api_key_here",
111        "MISTRAL_API_KEY": "your_mistral_api_key_here"
112      }
113    }
114  }
115}
116```
117 
118Restart Claude Desktop after configuration.
119 
120## Configuration
121 
122### API Keys
123 
124Set environment variables for the LLM providers you want to use:
125 
126```bash
127export OPENAI_API_KEY="your_openai_api_key"
128export ANTHROPIC_API_KEY="your_anthropic_api_key"
129export DEEPSEEK_API_KEY="your_deepseek_api_key"
130export GEMINI_API_KEY="your_gemini_api_key"
131export XAI_API_KEY="your_grok_api_key"
132export KIMI_API_KEY="your_kimi_api_key"
133export PERPLEXITY_API_KEY="your_perplexity_api_key"
134export MISTRAL_API_KEY="your_mistral_api_key"
135```
136 
137### Getting API Keys
138 
139- **OpenAI**: [https://platform.openai.com/api-keys](https://platform.openai.com/api-keys)
140- **Anthropic**: [https://console.anthropic.com/](https://console.anthropic.com/)
141- **DeepSeek**: [https://platform.deepseek.com/](https://platform.deepseek.com/)
142- **Google Gemini**: [https://makersuite.google.com/app/apikey](https://makersuite.google.com/app/apikey)
143- **xAI Grok**: [https://console.x.ai/](https://console.x.ai/)
144- **Moonshot AI**: [https://platform.moonshot.ai/](https://platform.moonshot.ai/)
145- **Perplexity**: [https://www.perplexity.ai/hub](https://www.perplexity.ai/hub)
146- **Mistral**: [https://console.mistral.ai/](https://console.mistral.ai/)
147 
148## Usage Examples
149 
150### Call ChatGPT
151 
152Get a response from OpenAI:
153 
154```json
155{
156  "tool": "call-chatgpt",
157  "arguments": {
158    "prompt": "Explain quantum computing in simple terms",
159    "temperature": 0.7,
160    "max_tokens": 500
161  }
162}
163```
164 
165### Call All LLMs
166 
167Get responses from all providers:
168 
169```json
170{
171  "tool": "call-all-llms",
172  "arguments": {
173    "prompt": "Write a short poem about AI",
174    "temperature": 0.8
175  }
176}
177```
178 
179### Set Tag-Based Preferences
180 
181Automatically use the best model for each task type:
182 
183```json
184{
185  "tool": "set-user-preferences",
186  "arguments": {
187    "defaultModel": "gpt-4o",
188    "costPreference": "cheaper",
189    "tagPreferences": {
190      "coding": "deepseek-r1",
191      "general": "gpt-4o",
192      "business": "claude-3.5-sonnet-20241022",
193      "reasoning": "deepseek-r1",
194      "math": "deepseek-r1",
195      "creative": "gpt-4o"
196    }
197  }
198}
199```
200 
201### Get Prompt History
202 
203View your prompt logs:
204 
205```json
206{
207  "tool": "get-prompt-history",
208  "arguments": {
209    "provider": "chatgpt",
210    "limit": 10
211  }
212}
213```
214 
215## Model Tags
216 
217Models are tagged by their strengths:
218 
219- **coding**: `deepseek-r1`, `deepseek-coder`, `gpt-4o`, `claude-3.5-sonnet-20241022`
220- **business**: `claude-3-opus-20240229`, `gpt-4o`, `gemini-1.5-pro`
221- **reasoning**: `deepseek-r1`, `o1-preview`, `claude-3.5-sonnet-20241022`
222- **math**: `deepseek-r1`, `o1-preview`, `o1-mini`
223- **creative**: `gpt-4o`, `claude-3-opus-20240229`, `gemini-1.5-pro`
224- **general**: `gpt-4o-mini`, `claude-3-haiku-20240307`, `gemini-1.5-flash`
225 
226## Use Cases
227 
228- **Multi-Perspective Analysis** – Get different perspectives from multiple LLMs
229- **Model Comparison** – Compare responses to understand strengths and weaknesses
230- **Cost Optimization** – Choose the most cost-effective model for each task
231- **Quality Assurance** – Cross-reference responses from multiple models
232- **Intelligent Selection** – Automatically use the best model for coding, business, reasoning, etc.
233- **Prompt Analytics** – Track usage, costs, and patterns with automatic logging
234 
235## Technical Details
236 
237**Built with:** Node.js, TypeScript, MCP SDK  
238**Dependencies:** `@modelcontextprotocol/sdk`, `superagent`, `zod`  
239**Platforms:** macOS, Windows, Linux
240 
241**Preference Storage:**
242 
243- Unix/macOS: `~/.cross-llm-mcp/preferences.json`
244- Windows: `%APPDATA%/cross-llm-mcp/preferences.json`
245 
246**Prompt Log Storage:**
247 
248- Unix/macOS: `~/.cross-llm-mcp/prompts.json`
249- Windows: `%APPDATA%/cross-llm-mcp/prompts.json`
250 
251## Contributing
252 
253⭐ **If this project helps you, please star it on GitHub!** ⭐
254 
255Contributions welcome! Please open an issue or submit a pull request.
256 
257## License
258 
259MIT License – see [LICENSE.md](LICENSE.md) for details.
260 
261## Support
262 
263If you find this project useful, consider supporting it:
264 
265**⚡ Lightning Network**
266 
267```
268lnbc1pjhhsqepp5mjgwnvg0z53shm22hfe9us289lnaqkwv8rn2s0rtekg5vvj56xnqdqqcqzzsxqyz5vqsp5gu6vh9hyp94c7t3tkpqrp2r059t4vrw7ps78a4n0a2u52678c7yq9qyyssq7zcferywka50wcy75skjfrdrk930cuyx24rg55cwfuzxs49rc9c53mpz6zug5y2544pt8y9jflnq0ltlha26ed846jh0y7n4gm8jd3qqaautqa
269```
270 
271**₿ Bitcoin**: [bc1ptzvr93pn959xq4et6sqzpfnkk2args22ewv5u2th4ps7hshfaqrshe0xtp](https://mempool.space/address/bc1ptzvr93pn959xq4et6sqzpfnkk2args22ewv5u2th4ps7hshfaqrshe0xtp)
272 
273**Ξ Ethereum/EVM**: [0x42ea529282DDE0AA87B42d9E83316eb23FE62c3f](https://etherscan.io/address/0x42ea529282DDE0AA87B42d9E83316eb23FE62c3f)
274