Configuration

Effero is configured via an effero.yaml file in your project root.

Top-level structure

agent:
  name: my-agent
  model:
    backend: litert          # litert | cactus | ollama | openai | anthropic | google
    model: /models/gemma-3-1b.litertlm
    device: npu              # auto | cpu | gpu | npu (for LiteRT-LM)
    fallback: [openai:gpt-4o] # ordered list of fallback backends
    min_confidence: 0.70     # confidence threshold for hybrid escalation
    hybrid_cloud_fallback: true # escalate low-confidence local runs to cloud

perception:
  audio:
    wake_word: "hey effero"
    asr: faster-whisper:small.en
    tts: piper:en_US-amy-medium
  vision:
    enabled: false
    backend: yolov9

skills:
  - iot.lights
  - iot.thermostat
  - computer_use.shell

safety:
  policy: safety/policies/home.yaml
  require_approval_for: []

CLI configuration

You can modify config from the terminal without editing YAML manually:

effero config set model.backend ollama --model qwen3:8b
effero config get model.backend

Environment variables

API keys for cloud LLM backends are loaded from environment variables:

Variable	Backend
OPENAI_API_KEY	OpenAI
ANTHROPIC_API_KEY	Anthropic (Claude)
GOOGLE_API_KEY	Google (Gemini)

Optional extras

Effero's pip install effero is deliberately lightweight. Enable additional capabilities with extras:

pip install effero[voice]    # ASR/TTS (faster-whisper, piper-tts)
pip install effero[vision]   # Detection (ultralytics/YOLO)
pip install effero[iot]      # MQTT/Matter protocols
pip install effero[litert]   # Google AI Edge LiteRT-LM in-process engine
pip install effero[llm]      # Cloud LLM SDKs (OpenAI, Anthropic, Google)
pip install effero[all]      # Everything above