config.ollama.yaml.example

# Development configuration for AWS MCP Server with Ollama (Local LLM)
server:
  port: 3000
  host: "localhost"

aws:
  region: "us-west-2"

mcp:
  server_name: "aws-infrastructure-server"
  version: "1.0.0"

agent:
  provider: "ollama"          # openai, gemini, anthropic, bedrock, ollama
  # Ollama server URL (default: http://localhost:11434)
  # You can set this via environment variable: OLLAMA_SERVER_URL
  # ollama_server_url: "http://localhost:11434"
  
  # Available Ollama models (examples - depends on what you have pulled):
  # Latest Models (2025):
  # - gemma3                  (Gemma 3 - 4B parameters, default)
  # - gemma3:1b               (Gemma 3 - 1B, 815MB, fastest)
  # - gemma3:12b              (Gemma 3 - 12B, 8.1GB)
  # - gemma3:27b              (Gemma 3 - 27B, 17GB)
  # - qwq                     (QwQ - 32B, 20GB, reasoning model)
  # - deepseek-r1             (DeepSeek R1 - 7B, 4.7GB, reasoning)
  # - deepseek-r1:671b        (DeepSeek R1 - 671B, 404GB, advanced reasoning)
  # 
  # Llama Models:
  # - llama4:scout            (Llama 4 Scout - 109B, 67GB)
  # - llama4:maverick         (Llama 4 Maverick - 400B, 245GB)
  # - llama3.3                (Llama 3.3 - 70B, 43GB)
  # - llama3.2                (Llama 3.2 - 3B, 2.0GB)
  # - llama3.2:1b             (Llama 3.2 - 1B, 1.3GB, faster)
  # - llama3.2-vision         (Llama 3.2 Vision - 11B, 7.9GB)
  # - llama3.2-vision:90b     (Llama 3.2 Vision - 90B, 55GB)
  # - llama3.1                (Llama 3.1 - 8B, 4.7GB)
  # - llama3.1:405b           (Llama 3.1 - 405B, 231GB)
  # 
  # Other Models:
  # - phi4                    (Phi 4 - 14B, 9.1GB)
  # - phi4-mini               (Phi 4 Mini - 3.8B, 2.5GB)
  # - mistral                 (Mistral - 7B, 4.1GB)
  # - codellama               (Code Llama - 7B, 3.8GB, code-focused)
  # - granite3.3              (Granite 3.3 - 8B, 4.9GB)
  # 
  # Note: You must pull the model first using: ollama pull <model-name>
  # Recommended: gemma3:27b (good balance) or deepseek-r1 (reasoning tasks)
  model: "gemma3:27b"
  
  # Context window size (number of tokens)
  # Larger values allow more context but use more memory
  # Typical values: 2048, 4096, 8192, 16384, 32768
  max_tokens: 8192
  
  temperature: 0.1
  dry_run: true
  auto_resolve_conflicts: false

logging:
  level: "debug"
  format: "text"
  output: "stdout"

state:
  file_path: "./states/infrastructure-state.json"
  backup_enabled: true
  backup_dir: "./backups"

web:
  port: 8080
  host: "localhost"
  template_dir: "web/templates"
  static_dir: "web/static"
  enable_websockets: true