Model Routing Playbook

Build a production routing policy across providers with fallback logic, cost caps, and quality gates.

0 uses 0 likes 5 views

System Prompt

Design a model routing playbook for {{use_case}}.
Primary model: @{{primary_provider}}.{{primary_model}}
Fallback model: @{{fallback_provider}}.{{fallback_model}}
Budget target (credits): {{budget_target}}
Latency SLA (ms): {{latency_sla_ms}}
Quality bar: {{quality_bar}}
Traffic pattern: {{traffic_pattern}}

Return JSON with keys: routing_rules, failover_rules, prompt_templates, monitoring_checks, and rollout_plan.

Variables

{{use_case}} (text) Required

Default: customer support copilot

{{primary_provider}} (text) Required

Default: openai

{{primary_model}} (text) Required

Default: gpt-5-mini

{{fallback_provider}} (text) Required

Default: anthropic

{{fallback_model}} (text) Required

Default: claude-3-5-sonnet

{{budget_target}} (number) Required

Default: 18

{{latency_sla_ms}} (number) Required

Default: 1800

{{quality_bar}} (text) Required

Default: >= 4.5/5 evaluator score

{{traffic_pattern}} (text)

Default: weekday spikes during business hours

Output Schema

{
  "kind": "ASK_TEMPLATE",
  "engine": "askgl",
  "askSlug": "demo_ava_brooks-blueprint",
  "variables": [
    {
      "name": "use_case",
      "type": "text",
      "label": "Use Case",
      "required": true,
      "defaultValue": "customer support copilot"
    },
    {
      "name": "primary_provider",
      "type": "select",
      "label": "Primary Provider",
      "options": [
        {
          "label": "OpenAI",
          "value": "openai"
        },
        {
          "label": "Anthropic",
          "value": "anthropic"
        },
        {
          "label": "Google",
          "value": "google"
        }
      ],
      "required": true,
      "defaultValue": "openai"
    },
    {
      "name": "primary_model",
      "type": "text",
      "label": "Primary Model",
      "required": true,
      "defaultValue": "gpt-5-mini"
    },
    {
      "name": "fallback_provider",
      "type": "select",
      "label": "Fallback Provider",
      "options": [
        {
          "label": "Anthropic",
          "value": "anthropic"
        },
        {
          "label": "OpenAI",
          "value": "openai"
        },
        {
          "label": "Google",
          "value": "google"
        }
      ],
      "required": true,
      "defaultValue": "anthropic"
    },
    {
      "name": "fallback_model",
      "type": "text",
      "label": "Fallback Model",
      "required": true,
      "defaultValue": "claude-3-5-sonnet"
    },
    {
      "name": "budget_target",
      "type": "number",
      "label": "Budget Target",
      "required": true,
      "defaultValue": 18
    },
    {
      "name": "latency_sla_ms",
      "type": "number",
      "label": "Latency SLA (ms)",
      "required": true,
      "defaultValue": 1800
    },
    {
      "name": "quality_bar",
      "type": "text",
      "label": "Quality Bar",
      "required": true,
      "defaultValue": ">= 4.5/5 evaluator score"
    },
    {
      "name": "traffic_pattern",
      "type": "text",
      "label": "Traffic Pattern",
      "required": false,
      "defaultValue": "weekday spikes during business hours"
    }
  ],
  "outputType": "json",
  "queryTemplate": "Design a model routing playbook for {{use_case}}.\nPrimary model: @{{primary_provider}}.{{primary_model}}\nFallback model: @{{fallback_provider}}.{{fallback_model}}\nBudget target (credits): {{budget_target}}\nLatency SLA (ms): {{latency_sla_ms}}\nQuality bar: {{quality_bar}}\nTraffic pattern: {{traffic_pattern}}\n\nReturn JSON with keys: routing_rules, failover_rules, prompt_templates, monitoring_checks, and rollout_plan."
}

Use This Prompt Fork & Customize