A3M Router - 90 Second Demo

#!/bin/bash

# Your LLM app today

const response = await openai.chat.completions.create({

model: "gpt-4o", // Always expensive

messages: [{ role: "user", content: "Explain quantum entanglement" }]

});

❌ Problem: $0.015 per request
❌ No fallback if OpenAI goes down
❌ Manual model selection forever

# With A3M: one line change

const response = await fetch("http://localhost:8787/v1/chat/completions", {

method: "POST",

headers: { "Content-Type": "application/json" },

body: JSON.stringify({

model: "auto", // ← A3M picks the best

messages: [{ role: "user", content: "Explain quantum entanglement" }]

})

});

📊 Task: explanation → Complexity: simple → Budget: minimize
✅ Routed to: Groq (FREE, 847ms)
   Cost: $0.00 | Quality: 94%

# A3M gracefully handles provider failures...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ [Groq] — Attempting connection... [Groq] — ✗ FAILED — 503 Service Unavailable [Circuit Breaker] — Tripped after 3 failures [DeepSeek] — HEALTHY — Switching... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

✅ Response delivered via DeepSeek fallback
   Your app: never knew there was a problem

# Same query. Dramatically different costs.

WITHOUT A3M

$0.0015

GPT-4o @ $3.00/1K

500 tokens

WITH A3M

$0.00

Groq @ $0.00/1K

500 tokens

💰 SAVINGS: 100% per query
   At 1000 queries/day → $1.50 saved daily
   At 1000 queries/day → $547 saved yearly

Get started in 10 seconds:

npm install -g adaptive-memory-multi-model-router

# Auto-detects your API keys, zero config

npx a3m-router serve

# Change: model: 'gpt-4o' → model: 'auto'

🔗 github.com/Das-rebel/adaptive-memory-multi-model-router
📦 npmjs.com/package/adaptive-memory-multi-model-router

✨ One prompt in. The right model out. ✨

🔀 A3M Router