mcp-claude by NimbleBrainInc - MCP Server

Anthropic Claude MCP Server

MCP server providing comprehensive access to Anthropic's Claude API.

Features

Core Capabilities

Messages API - Full access to Claude 3.5 Sonnet, Haiku, and Claude 3 Opus
Simple Chat - Easy-to-use chat interface
Vision - Analyze images via URL or base64
Multi-turn Conversations - Maintain conversation history
Token Counting - Estimate token usage before making calls
Model Comparison - Compare responses across different Claude models
Structured Extraction - Extract structured data from unstructured text
Model Information - Get details about Claude models and capabilities

Installation

Clone this repository
Install dependencies:

pip install -r requirements.txt

Set up your environment variables:

cp .env.example .env
# Edit .env and add your Anthropic API key

Configuration

Get Your Anthropic API Key

Go to https://console.anthropic.com/settings/keys
Sign in or create an account
Click "Create Key"
Copy the key and add it to your .env file

Claude Desktop Configuration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "claude": {
      "command": "fastmcp",
      "args": ["run", "server.py"],
      "env": {
        "ANTHROPIC_API_KEY": "your_api_key_here"
      }
    }
  }
}

Available Tools

chat

Simple chat interface - send a prompt, get a response.

Parameters:

prompt (required): Your message or question
model: Claude model (default: "claude-3-5-sonnet-20241022")
max_tokens: Maximum response length (default: 4096)
system: Optional system prompt for context
temperature: Creativity level 0-1

Example:

{
  "prompt": "Explain quantum computing in simple terms",
  "max_tokens": 500
}

create_message

Full Messages API with all parameters and features.

Parameters:

messages (required): Conversation history
model: Claude model
max_tokens (required): Max response length
system: System prompt
temperature, top_p, top_k: Sampling parameters
stop_sequences: Custom stop sequences

analyze_image

Analyze images using Claude's vision capabilities.

Parameters:

image_url (required): URL of image to analyze
prompt: Question about the image (default: "What's in this image?")
model: Vision-capable model
max_tokens: Max response length

Example:

{
  "image_url": "https://example.com/photo.jpg",
  "prompt": "Describe the architectural style of this building"
}

analyze_image_base64

Analyze base64-encoded images.

Parameters:

image_base64 (required): Base64 image data
media_type (required): image/jpeg, image/png, image/gif, or image/webp
prompt: Question about the image
model: Vision model
max_tokens: Max response length

multi_turn_conversation

Continue a conversation with context from previous messages.

Parameters:

conversation_history (required): Previous messages
new_message (required): New user message
model: Claude model
max_tokens: Max response length
system: Optional system prompt

Returns: Response plus updated conversation history

Example:

{
  "conversation_history": [
    {"role": "user", "content": "What's the capital of France?"},
    {"role": "assistant", "content": "The capital of France is Paris."}
  ],
  "new_message": "What's the population?"
}

count_tokens

Estimate token usage for messages.

Parameters:

messages (required): Messages to count tokens for
model: Model (affects tokenization)
system: Optional system prompt

Returns: Input and output token counts

get_model_info

Get information about Claude models.

Parameters:

model: Model identifier (default: "claude-3-5-sonnet-20241022")

Returns: Context window, capabilities, tier, description

compare_responses

Get responses from multiple models for comparison.

Parameters:

prompt (required): Question or task
models: List of models to compare (max 3, default: Sonnet and Haiku)
max_tokens: Max tokens per response
system: Optional system prompt

Returns: Responses from each model with usage stats

Example:

{
  "prompt": "Write a haiku about programming",
  "models": ["claude-3-5-sonnet-20241022", "claude-3-5-haiku-20241022"]
}

extract_structured_data

Extract structured data from unstructured text.

Parameters:

text (required): Unstructured text
schema_description (required): Description of desired structure
model: Claude model
max_tokens: Max response length

Example:

{
  "text": "Contact John Smith at john@example.com or call 555-1234",
  "schema_description": "Extract as JSON with fields: name, email, phone"
}

Claude Models

Claude 3.5 Family

claude-3-5-sonnet-20241022: Most intelligent, best overall performance
claude-3-5-haiku-20241022: Fastest model, cost-effective

Claude 3 Family

claude-3-opus-20240229: Top-tier performance for complex tasks
claude-3-sonnet-20240229: Balanced intelligence and speed
claude-3-haiku-20240307: Fast and economical

All models support:

200K context window
Vision capabilities
Tool use
Multi-turn conversations

Usage Examples

Simple Question

{
  "tool": "chat",
  "prompt": "What are the three laws of robotics?"
}

Creative Writing with System Prompt

{
  "tool": "chat",
  "prompt": "Write a short story about a time traveler",
  "system": "You are a creative writer specializing in science fiction",
  "temperature": 0.8,
  "max_tokens": 1000
}

Image Analysis

{
  "tool": "analyze_image",
  "image_url": "https://example.com/diagram.png",
  "prompt": "Explain what this technical diagram shows"
}

Ongoing Conversation

{
  "tool": "multi_turn_conversation",
  "conversation_history": [
    {"role": "user", "content": "I'm planning a trip to Japan"},
    {"role": "assistant", "content": "That sounds exciting! How long will you be staying?"}
  ],
  "new_message": "Two weeks in spring"
}

Model Comparison

{
  "tool": "compare_responses",
  "prompt": "Explain the theory of relativity",
  "models": ["claude-3-5-sonnet-20241022", "claude-3-opus-20240229"]
}

Data Extraction

{
  "tool": "extract_structured_data",
  "text": "Apple Inc. reported revenue of $394.3B in fiscal 2024, up 2% YoY",
  "schema_description": "Extract as JSON: company, revenue, year, growth_rate"
}

Error Handling

The server includes comprehensive error handling for:

Invalid API keys
Rate limiting (429 errors)
Token limit exceeded (400 errors)
Invalid model names
Malformed requests
Network timeouts

Rate Limits

Anthropic enforces rate limits based on your account tier:

Free tier: Limited requests per day
Build tier: Higher daily limits
Scale tier: Enterprise-level limits

Check your limits at: https://console.anthropic.com/settings/limits

Pricing

Approximate costs per million tokens (input/output):

Claude 3.5 Sonnet: $3 / $15
Claude 3.5 Haiku: $0.80 / $4
Claude 3 Opus: $15 / $75
Claude 3 Sonnet: $3 / $15
Claude 3 Haiku: $0.25 / $1.25

Current pricing: https://www.anthropic.com/pricing

Best Practices

Use System Prompts: Set context and instructions for better responses
Manage Context: Keep conversation history relevant, trim old messages
Choose the Right Model: Haiku for speed, Sonnet for balance, Opus for complexity
Token Management: Use count_tokens to estimate costs before calls
Error Handling: Implement retry logic for rate limits
Vision: Images count toward token limits (analyze_image tools)

Security Notes

Never commit your API key to version control
Use environment variables for sensitive data
Rotate API keys regularly
Monitor usage in Anthropic Console
Set spending limits in your account

NimbleBrainInc/mcp-claude

Anthropic Claude MCP Server

Features

Core Capabilities

Installation

Configuration

Get Your Anthropic API Key

Claude Desktop Configuration

Available Tools

chat

create_message

analyze_image

analyze_image_base64

multi_turn_conversation

count_tokens

get_model_info

compare_responses

extract_structured_data

Claude Models

Claude 3.5 Family

Claude 3 Family

Usage Examples

Simple Question

Creative Writing with System Prompt

Image Analysis

Ongoing Conversation

Model Comparison

Data Extraction

Error Handling

Rate Limits

Pricing

Best Practices

Security Notes

Troubleshooting

"Invalid API Key" Error

Rate Limit Errors (429)

Context Length Errors

Resources

License