Skip to main content
Glama

MCP AI Hub

by feiskyer
config_example.yaml7.97 kB
# MCP AI Hub Configuration File # ============================== # # This is a comprehensive example configuration for MCP AI Hub. # Copy this file to ~/.ai_hub.yaml and customize it for your needs. # # IMPORTANT SECURITY NOTES: # - Replace ALL placeholder API keys with your actual keys # - Set file permissions: chmod 600 ~/.ai_hub.yaml # - NEVER commit this file to version control # - Use environment variables or secret management for production # Global System Prompt (Optional) # -------------------------------- # This system prompt will be applied to ALL models unless overridden by a model-specific prompt. # Useful for setting consistent behavior across all AI interactions. # Removal tip: Comment out or delete this line to disable global system prompts. global_system_prompt: "You are a helpful AI assistant. Be concise and accurate in your responses." # Model Configuration # ------------------- # Each entry in model_list defines a model you can use with MCP AI Hub # The model_name is what you'll use in MCP tools (e.g., chat("gpt-4", "Hello")) # The litellm_params configure the actual provider connection # Optional: Add system_prompt field to override global system prompt for specific models model_list: # ============================================================================ # OPENAI MODELS # ============================================================================ - model_name: gpt-4o # Friendly name for MCP tools # Model-specific system prompt (overrides global_system_prompt for this model only) system_prompt: "You are GPT-4o, OpenAI's most advanced multimodal model. You can process text, images, and other inputs." litellm_params: # LiteLM provider/model identifier (format: provider/model-name) model: openai/gpt-4o # Your actual OpenAI API key - REPLACE THIS! api_key: "sk-your-openai-api-key-here" # Maximum tokens in response (optional, provider-specific limits apply) max_tokens: 4096 # Response creativity: 0.0 (deterministic) to 1.0 (creative) (optional) temperature: 0.7 # Additional parameters you can use: # top_p: 0.9 # Nucleus sampling parameter # frequency_penalty: 0.1 # Reduce repetition # presence_penalty: 0.1 # Encourage topic diversity # stop: ["\n\n", "Human:"] # Stop sequences - model_name: gpt-5 litellm_params: model: openai/gpt-5 api_key: "sk-your-openai-api-key-here" # REPLACE WITH YOUR KEY max_tokens: 4096 # Higher token limit for longer responses temperature: 0.7 # ============================================================================ # ANTHROPIC MODELS # ============================================================================ - model_name: claude-sonnet # Claude 4 Sonnet - balanced performance # Model-specific system prompt (overrides global_system_prompt for this model only) system_prompt: "You are Claude 4 Sonnet, an AI assistant created by Anthropic. You excel at coding, analysis, and creative tasks." litellm_params: # Full model identifier with version model: anthropic/claude-sonnet-4-20250514 # Your actual Anthropic API key - REPLACE THIS! api_key: "sk-ant-your-anthropic-api-key-here" max_tokens: 12000 temperature: 0.7 # ============================================================================ # GOOGLE MODELS # ============================================================================ - model_name: gemini-2.5-pro litellm_params: model: gemini/gemini-2.5-pro api_key: "your-gemini-api-key" # REPLACE WITH YOUR KEY max_tokens: 8000 temperature: 0.7 - model_name: gemini-2.5-flash-image-preview litellm_params: model: gemini/gemini-2.5-flash-image-preview api_key: "your-gemini-api-key" # REPLACE WITH YOUR KEY max_tokens: 8000 temperature: 0.7 # ============================================================================ # AZURE OPENAI MODELS # ============================================================================ # Uncomment and configure if you use Azure OpenAI Service # - model_name: azure-gpt4 # litellm_params: # model: azure/gpt-4 # api_key: "your-azure-api-key-here" # REPLACE WITH YOUR KEY # # Azure-specific: your Azure OpenAI endpoint # api_base: "https://your-resource-name.openai.azure.com/" # # Azure-specific: API version # api_version: "2024-02-15-preview" # max_tokens: 2048 # temperature: 0.7 # ============================================================================ # AWS BEDROCK MODELS # ============================================================================ # Uncomment and configure if you use AWS Bedrock # - model_name: bedrock-claude # litellm_params: # model: bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0 # # AWS credentials (can also use environment variables) # aws_access_key_id: "your-aws-access-key" # aws_secret_access_key: "your-aws-secret-key" # aws_region_name: "us-east-1" # max_tokens: 4096 # temperature: 0.7 # ============================================================================ # TOGETHER AI MODELS (Open Source) # ============================================================================ # Uncomment and configure if you use Together AI # - model_name: llama-70b # litellm_params: # model: together_ai/meta-llama/Llama-2-70b-chat-hf # api_key: "your-together-api-key-here" # REPLACE WITH YOUR KEY # max_tokens: 2048 # temperature: 0.7 # - model_name: mistral-7b # litellm_params: # model: together_ai/mistralai/Mistral-7B-v0.1 # api_key: "your-together-api-key-here" # REPLACE WITH YOUR KEY # max_tokens: 2048 # temperature: 0.7 # ============================================================================ # HUGGING FACE MODELS # ============================================================================ # Uncomment and configure if you use Hugging Face Inference API # - model_name: hf-model # litellm_params: # model: huggingface/your-model-name # api_key: "your-hf-api-key-here" # REPLACE WITH YOUR KEY # max_tokens: 2048 # temperature: 0.7 # ============================================================================ # CUSTOM ENDPOINTS AND PROXY CONFIGURATION # ============================================================================ # Use these examples for custom setups, proxy servers, or local deployments # Example: Corporate proxy server # - model_name: gpt-4-corporate # litellm_params: # model: openai/gpt-4 # api_key: "sk-your-api-key" # REPLACE WITH YOUR KEY # # Custom OpenAI-compatible endpoint # api_base: "https://corporate-proxy.example.com/v1" # max_tokens: 2048 # temperature: 0.7 # Example: Local LLM server (Ollama, LM Studio, vLLM, etc.) # - model_name: local-llama # litellm_params: # # Local servers often use OpenAI-compatible format # model: openai/llama-2-7b-chat # # Local servers often accept any API key # api_key: "dummy-key" # # Your local server endpoint # api_base: "http://localhost:8080/v1" # max_tokens: 2048 # temperature: 0.7 # Example: Custom Anthropic endpoint # - model_name: claude-custom # litellm_params: # model: anthropic/claude-3-5-sonnet-20241022 # api_key: "sk-ant-your-api-key" # REPLACE WITH YOUR KEY # # Custom Anthropic-compatible endpoint # api_base: "https://custom-anthropic.example.com/v1" # max_tokens: 4096 # temperature: 0.7 # ============================================================================ # ADVANCED CONFIGURATION OPTIONS # ============================================================================ # For more providers, please refer to the LiteLLM docs: https://docs.litellm.ai/docs/providers

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/feiskyer/mcp-ai-hub'

If you have feedback or need assistance with the MCP directory API, please join our Discord server