ask-ai-mcp

Ask AI MCP Server

A Model Context Protocol (MCP) server that enables AI models to consult with another AI model through tool calling. This server acts as a bridge, allowing AI assistants to ask questions to different AI models when they need help, alternative perspectives, code reviews, or expert opinions on complex problems. It supports multiple AI providers and is designed for seamless integration with MCP-compatible clients.

✨ Features

• Multiple AI Providers: OpenAI, Anthropic, Google AI Studio, Perplexity, and OpenAI-compatible endpoints
• Reasoning Model Support: Configurable reasoning effort for advanced models across all providers
• Efficient Context Handling: Single-call design for maximum efficiency and cost-effectiveness
• Flexible Configuration: Optional parameters for temperature, token limits, and custom endpoints
• Production Ready: Comprehensive error handling, validation, and TypeScript support

📦 Installation

Prerequisites

• Node.js 24.0.0 or higher
• API key for your chosen AI provider

Quick Start (Recommended)

The easiest way to use this MCP server is with npx:

{
  "mcpServers": {
    "ask-ai": {
      "command": "npx",
      "args": ["ask-ai-mcp"],
      "env": {
        "PROVIDER": "openai",
        "MODEL": "o3-2025-04-16",
        "API_KEY": "your_api_key_here"
      }
    }
  }
}

From Source (Advanced)

For development or customization:

git clone <link to repository>
cd ask-ai-mcp
pnpm install
pnpm run build

Then reference the built server in your MCP configuration:

{
  "mcpServers": {
    "ask-ai": {
      "command": "node",
      "args": ["/path/to/ask-ai-mcp/dist/server.js"],
      "env": {
        "PROVIDER": "openai",
        "MODEL": "o3-2025-04-16",
        "API_KEY": "your_api_key_here"
      }
    }
  }
}

⚙️ Configuration

Required Environment Variables

Variable	Description
PROVIDER	The AI provider to use. openai, anthropic, google, perplexity, openai-compatible
MODEL	The model name from the selected provider.
API_KEY	Your API key for the chosen provider.

Optional Environment Variables

Variable	Description
TEMPERATURE	Sampling temperature for the model (e.g., 0.7).
MAX_TOKENS	Maximum number of tokens to generate (e.g., 10000).
BASE_URL	Custom endpoint for openai-compatible providers.
REASONING_EFFORT	For reasoning models: low, medium, high, max, or a budget string (e.g., "10000").

Provider Examples

OpenAI

PROVIDER=openai
MODEL=o3-2025-04-16
API_KEY=your_openai_api_key
# Optional
REASONING_EFFORT=medium  # low, medium, high, max, or budget like "10000"

Anthropic

PROVIDER=anthropic
MODEL=claude-opus-4-20250514
API_KEY=your_anthropic_api_key

Google AI Studio

PROVIDER=google
MODEL=gemini-2.5-pro-preview-06-05
API_KEY=your_google_api_key

Perplexity

PROVIDER=perplexity
MODEL=sonar-pro
API_KEY=your_perplexity_api_key

OpenAI-Compatible

PROVIDER=openai-compatible
MODEL=your_model_name
API_KEY=your_api_key
BASE_URL=https://your-endpoint.com/v1

🛠️ Available Tool

Once integrated, AI models can use the ask_ai tool:

{
  "name": "ask_ai",
  "arguments": {
    "question": "How can I optimize this React component?",
    "context": "I have a React component that renders 1000+ items and re-renders frequently causing performance issues. Current code: [include your code here]"
  }
}

📄 License

This project is licensed under the GNU General Public License, version 2.