model-router-policy-py 0.1.0

Policy-based model routing by capability, cost, latency, and privacy. Python port of @mukundakatta/model-router-policy.

model-router-policy-py

Policy-based model routing by capability, cost, latency, and privacy. Pure Python, zero deps. Python port of @mukundakatta/model-router-policy.

pip install model-router-policy-py

from model_router_policy import route_model

models = [
    {"name": "haiku",  "capabilities": ["chat"],          "quality": 0.7,  "cost_per_1k_tokens": 0.5,  "p50_latency_ms": 400,  "allows_private_data": True},
    {"name": "sonnet", "capabilities": ["chat", "tools"], "quality": 0.9,  "cost_per_1k_tokens": 3,    "p50_latency_ms": 800,  "allows_private_data": True},
    {"name": "opus",   "capabilities": ["chat", "tools", "vision"], "quality": 0.95, "cost_per_1k_tokens": 15, "p50_latency_ms": 1500, "allows_private_data": False},
]

route_model({"required_capabilities": ["vision"], "private_data": True}, models)
# RoutingDecision(model=None, reason='no_model_satisfies_policy', ...)  # opus excluded by privacy

Score weighting: quality 0.55 + cost 0.25 + latency 0.20 + preferred-cap fit 0.10/cap.

License

MIT