← All recipes

LLM Ensemble (capacity-weighted)

aiintermediate

Route completions across multiple model runners; weights follow spare capacity and price.

CapacityRegistryBackpressurePoolPricingCurveEscrowBufferCompletionTrackerStakeManager

How to deploy

# Install the CLI

npm install -g @backproto/bake

# Fork this recipe

bake fork llm-ensemble-router

# Simulate with fault injection

bake sim --chaos

# Deploy to Base Sepolia

bake deploy