Configuration

Shared runtime defaults live in maki/config.py. Every value can be overridden with an environment variable or a .env file (python-dotenv is supported).

Variable	Description
MAKI_OLLAMA_BASE_URL	Full Ollama base URL
MAKI_OLLAMA_HOST	Ollama hostname
MAKI_OLLAMA_PORT	Ollama port
MAKI_LLAMACPP_BASE_URL	llama.cpp server URL (default http://127.0.0.1:8080/v1)
MAKI_LLAMACPP_MODEL	llama.cpp model name (default local-model; set it in router mode)
LLAMACPP_API_KEY	llama.cpp API key, only if the server was started with --api-key
MAKI_DEFAULT_MODEL	Default model name
MAKI_DEFAULT_TEMPERATURE	Sampling temperature
MAKI_REQUEST_TIMEOUT	Per-request timeout (seconds)
MAKI_HTTP_TIMEOUT	Low-level HTTP timeout
MAKI_LOG_LEVEL	Logging level
MAKI_WEB_USER_AGENT	User-agent string for web plugins

Example .env for an Ollama server on another machine on your LAN:

MAKI_OLLAMA_BASE_URL=http://192.168.1.20:11434
MAKI_REQUEST_TIMEOUT=300

Logging

Maki does not configure logging on import. Call it explicitly:

from maki.logging_config import configure_logging

configure_logging()