adaptive-memory-multi-model-router 2.12.0 → 2.12.1

package/demo-simple.tape

@@ -0,0 +1,205 @@
+# A3M Router VHS Demo - 90 Second Terminal Demo
+# This version works without ttyd - outputs to GIF directly
+# Target: 60-90 seconds
+
+# =============================================================================
+# SETUP - Works with just image output
+# =============================================================================
+
+Output a3m-demo.gif
+Set FontSize 16
+Set Width 1200
+Set Height 675
+Set TypingSpeed 30ms
+
+# =============================================================================
+# SCENE 1: THE PROBLEM (0-15s)
+# =============================================================================
+
+Type "#!/bin/bash"
+Enter
+
+Type "# Your LLM app today — costs $0.015 per request, no fallback"
+Enter
+
+Type "cat << 'EOF'"
+Enter
+Type "// Without A3M: you manage everything manually"
+Enter
+Type "const response = await openai.chat.completions.create({"
+Enter
+Type "  model: \"gpt-4o\",  // Always expensive"
+Enter
+Type "  messages: [{ role: \"user\", content: \"Explain quantum entanglement\" }]"
+Enter
+Type "});"
+Enter
+Type "// If OpenAI goes down → your app crashes"
+Enter
+Type "// You pay $0.015 per request forever"
+Enter
+Type "EOF"
+Enter
+
+Sleep 1s
+Type "echo \"❌ Problem: $0.015/req, no fallback, manual management\""
+Enter
+
+Sleep 2s
+
+# =============================================================================
+# SCENE 2: WITH A3M - ONE LINE CHANGE (15-35s)
+# =============================================================================
+
+Ctrl+L
+
+Type "# With A3M: one line change, infinite benefits"
+Enter
+
+Type "cat << 'EOF'"
+Enter
+Type "// Same request with A3M router"
+Enter
+Type "const response = await fetch(\"http://localhost:8787/v1/chat/completions\", {"
+Enter
+Type "  method: \"POST\","
+Enter
+Type "  headers: { \"Content-Type\": \"application/json\" },"
+Enter
+Type "  body: JSON.stringify({"
+Enter
+Type "    model: \"auto\",  // ← A3M picks the best model"
+Enter
+Type "    messages: [{ role: \"user\", content: \"Explain quantum entanglement\" }]"
+Enter
+Type "  })"
+Enter
+Type "});"
+Enter
+Type "EOF"
+Enter
+
+Sleep 1s
+
+Type "# A3M analyzes and decides:"
+Enter
+Type "echo \"📊 Task: explanation → Complexity: simple → Budget: minimize\""
+Enter
+Type "echo \"✅ Routed to: Groq (FREE, 847ms)\""
+Enter
+Type "echo \"   Cost: $0.00 | Quality: 94%\""
+Enter
+
+Sleep 2s
+
+# =============================================================================
+# SCENE 3: FAILURE RECOVERY (35-55s)
+# =============================================================================
+
+Ctrl+L
+
+Type "# A3M gracefully handles provider failures..."
+Enter
+Sleep 1s
+
+Type "echo \"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\""
+Enter
+Type "echo \"  [Groq] — Attempting connection...\""
+Enter
+Sleep 1s
+Type "echo \"  [Groq] — ✗ FAILED — 503 Service Unavailable\""
+Enter
+Sleep 500ms
+Type "echo \"  [Circuit Breaker] — Tripped after 3 failures\""
+Enter
+Sleep 500ms
+Type "echo \"  [DeepSeek] — HEALTHY — Switching...\""
+Enter
+Sleep 1s
+Type "echo \"\""
+Enter
+Type "echo \"✅ Response delivered via DeepSeek fallback\""
+Enter
+Type "echo \"   Your app: never knew there was a problem\""
+Enter
+Type "echo \"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\""
+Enter
+
+Sleep 2s
+
+# =============================================================================
+# SCENE 4: COST SAVINGS (55-70s)
+# =============================================================================
+
+Ctrl+L
+
+Type "# Same query. Dramatically different costs."
+Enter
+
+Type "echo \"\""
+Enter
+Type "echo \"  WITHOUT A3M          WITH A3M\""
+Enter
+Type "echo \"  ─────────────────    ────────────────\""
+Enter
+Type "echo \"  GPT-4o @ $3.00/1K     Groq @ $0.00/1K\""
+Enter
+Type "echo \"  500 tokens = $0.0015  500 tokens = $0.00\""
+Enter
+Sleep 1s
+Type "echo \"\""
+Enter
+Type "echo \"💰 SAVINGS: 100% per query\""
+Enter
+Type "echo \"   At 1000 queries/day → $1.50 saved daily\""
+Enter
+Type "echo \"   At 1000 queries/day → $547 saved yearly\""
+Enter
+
+Sleep 2s
+
+# =============================================================================
+# SCENE 5: ONE COMMAND (70-90s)
+# =============================================================================
+
+Ctrl+L
+
+Type "echo \"\""
+Enter
+Type "echo \"╔═══════════════════════════════════════════════════╗\""
+Enter
+Type "echo \"║         Get started in 10 seconds:                ║\""
+Enter
+Type "echo \"╚═══════════════════════════════════════════════════╝\""
+Enter
+Sleep 500ms
+Type "echo \"\""
+Enter
+Type "echo \"  npm install -g adaptive-memory-multi-model-router\""
+Enter
+Sleep 300ms
+Type "echo \"\""
+Enter
+Type "echo \"  # Auto-detects your API keys, zero config\""
+Enter
+Type "echo \"  npx a3m-router serve\""
+Enter
+Sleep 500ms
+Type "echo \"\""
+Enter
+Type "echo \"  # Change: model: 'gpt-4o' → model: 'auto'\""
+Enter
+Sleep 500ms
+Type "echo \"\""
+Enter
+Type "echo \"  GitHub: github.com/Das-rebel/adaptive-memory-multi-model-router\""
+Enter
+Type "echo \"  npm:    npmjs.com/package/adaptive-memory-multi-model-router\""
+Enter
+Sleep 1s
+Type "echo \"\""
+Enter
+Type "echo \"✨ One prompt in. The right model out. ✨\""
+Enter
+
+Sleep 2s