claude-code-router-config 1.0.1 → 1.1.0

package/templates/README.md

@@ -0,0 +1,161 @@
+# Configuration Templates
+
+This directory contains pre-configured templates optimized for different use cases. Each template provides a complete configuration with optimized provider selection, routing rules, and settings.
+
+## Available Templates
+
+### 1. 🚀 `performance-optimized.json`
+**Best for**: Real-time applications, chatbots, time-sensitive tasks
+- **Priority**: Speed and low latency
+- **Expected Latency**: 500-2000ms
+- **Cost Profile**: Low to Medium
+- **Primary Provider**: Gemini Flash (fastest)
+- **Fallback**: OpenAI Mini, Qwen Turbo
+
+### 2. 💰 `cost-optimized.json`
+**Best for**: Budget-conscious users, bulk processing, non-critical tasks
+- **Priority**: Maximum cost savings
+- **Expected Savings**: 70-90% vs premium models
+- **Cost Profile**: Very Low
+- **Primary Provider**: Qwen Turbo (cheapest)
+- **Features**: Budget tracking and alerts
+
+### 3. 🎯 `quality-focused.json`
+**Best for**: Critical applications, complex reasoning, professional work
+- **Priority**: Highest quality and accuracy
+- **Expected Latency**: 3-10 seconds
+- **Cost Profile**: High
+- **Primary Provider**: Claude Sonnet 4 (best reasoning)
+- **Features**: Enhanced reasoning, chain-of-thought
+
+### 4. 💻 `development.json`
+**Best for**: Developers, coding assistance, software engineering
+- **Priority**: Code generation and debugging
+- **Specialized**: Coding models and frameworks
+- **Cost Profile**: Medium
+- **Primary Provider**: OpenAI GPT-4o Mini (best for coding)
+- **Features**: Code context awareness, project intelligence
+
+### 5. ⚖️ `balanced.json` **(Default)**
+**Best for**: General use, mixed workloads, everyday tasks
+- **Priority**: Balanced cost-performance-quality
+- **Expected Latency**: 2-4 seconds
+- **Cost Profile**: Medium
+- **Primary Provider**: OpenAI GPT-4o (versatile)
+- **Features**: Adaptive routing, auto-optimization
+
+## Usage
+
+### Apply a Template:
+```bash
+# Using CLI (recommended)
+ccr config template performance-optimized
+ccr config template cost-optimized
+ccr config template quality-focused
+ccr config template development
+ccr config template balanced
+
+# Manual copy
+cp templates/performance-optimized.json ~/.claude-code-router/config.json
+```
+
+### Compare Templates:
+```bash
+# See quick comparison
+ccr config compare --all
+
+# Compare specific templates
+ccr config compare performance-optimized cost-optimized
+```
+
+## Template Features
+
+### Performance Optimization
+- Request timeout settings
+- Retry policies
+- Connection pooling
+- Caching strategies
+
+### Cost Management
+- Budget limits and alerts
+- Cost tracking
+- Provider cost analysis
+- Usage optimization
+
+### Quality Enhancement
+- Temperature settings
+- Model-specific parameters
+- Prompt optimization
+- Response validation
+
+### Development Tools
+- Code context awareness
+- Language-specific optimization
+- Debug mode
+- Testing configurations
+
+## Customization
+
+Templates can be customized by:
+1. Copying a template as a base
+2. Modifying providers and models
+3. Adjusting routing rules
+4. Fine-tuning parameters
+5. Adding custom features
+
+Example customization:
+```json
+{
+  "customSetting": {
+    "myPreference": true,
+    "specificModel": "claude-sonnet-4-latest"
+  }
+}
+```
+
+## Switching Templates
+
+To switch between templates:
+```bash
+# Backup current config
+ccr config backup
+
+# Apply new template
+ccr config template quality-focused
+
+# Verify configuration
+ccr config validate
+
+# Test providers
+ccr benchmark full
+```
+
+## Recommendations
+
+| Use Case | Recommended Template |
+|----------|---------------------|
+| **Chat Applications** | performance-optimized |
+| **Data Processing** | cost-optimized |
+| **Research & Analysis** | quality-focused |
+| **Software Development** | development |
+| **General Purpose** | balanced |
+
+## Tips
+
+1. **Start with balanced** if unsure - it works well for most cases
+2. **Monitor costs** when using quality-focused templates
+3. **Test latency** with performance-optimized for your specific use case
+4. **Customize** templates based on your actual usage patterns
+5. **Backup** configurations before switching templates
+
+## Template Comparison
+
+| Feature | Performance | Cost | Quality | Development | Balanced |
+|---------|-------------|------|---------|--------------|----------|
+| **Speed** | ⭐⭐⭐⭐⭐ | ⭐⭐⭐ | ⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐ |
+| **Cost** | ⭐⭐⭐ | ⭐⭐⭐⭐⭐ | ⭐ | ⭐⭐⭐ | ⭐⭐⭐ |
+| **Quality** | ⭐⭐⭐ | ⭐⭐ | ⭐⭐⭐⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐ |
+| **Coding** | ⭐⭐⭐ | ⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐⭐ | ⭐⭐⭐⭐ |
+| **Best For** | Speed | Budget | Accuracy | Development | Everything |
+
+Choose the template that best matches your priorities and use case!

package/templates/balanced.json

@@ -0,0 +1,111 @@
+{
+  "LOG": true,
+  "LOG_LEVEL": "info",
+  "API_TIMEOUT_MS": 30000,
+  "CUSTOM_ROUTER_PATH": "$HOME/.claude-code-router/intent-router.js",
+
+  "Providers": [
+    {
+      "name": "openai",
+      "api_base_url": "https://api.openai.com/v1/chat/completions",
+      "api_key": "$OPENAI_API_KEY",
+      "models": ["gpt-4o", "gpt-4o-mini", "gpt-4-turbo"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 1,
+      "description": "Versatile performer for most tasks",
+      "strengths": ["coding", "reasoning", "general"]
+    },
+    {
+      "name": "anthropic",
+      "api_base_url": "https://api.anthropic.com/v1/messages",
+      "api_key": "$ANTHROPIC_API_KEY",
+      "models": ["claude-3-5-sonnet-latest", "claude-sonnet-4-latest"],
+      "transformer": {
+        "use": ["Anthropic"]
+      },
+      "priority": 2,
+      "description": "Excellent for analysis and complex reasoning",
+      "strengths": ["analysis", "writing", "reasoning"]
+    },
+    {
+      "name": "gemini",
+      "api_base_url": "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
+      "api_key": "$GEMINI_API_KEY",
+      "models": ["gemini-2.5-flash", "gemini-2.5-pro"],
+      "transformer": {
+        "use": ["gemini"]
+      },
+      "priority": 3,
+      "description": "Fast with large context and good quality",
+      "strengths": ["speed", "context", "multilingual"]
+    },
+    {
+      "name": "qwen",
+      "api_base_url": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
+      "api_key": "$QWEN_API_KEY",
+      "models": ["qwen-plus", "qwen-max"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 4,
+      "description": "Cost-effective with good performance",
+      "strengths": ["cost", "multilingual", "coding"]
+    },
+    {
+      "name": "glm",
+      "api_base_url": "https://api.z.ai/api/paas/v4/chat/completions",
+      "api_key": "$GLM_API_KEY",
+      "models": ["glm-4.6", "glm-4.5"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 5,
+      "description": "Specialized in Chinese and multilingual tasks",
+      "strengths": ["multilingual", "translation", "chinese"]
+    }
+  ],
+
+  "Router": {
+    "default": "openai,gpt-4o",
+    "background": "gemini,gemini-2.5-flash",
+    "think": "anthropic,claude-3-5-sonnet-latest",
+    "longContext": "gemini,gemini-2.5-pro",
+    "longContextThreshold": 100000,
+    "Balanced": {
+      "enabled": true,
+      "adaptiveRouting": true,
+      "costAware": true,
+      "performanceAware": true
+    }
+  },
+
+  "BalancedOptimization": {
+    "costQualityRatio": 0.7,
+    "performanceWeight": 0.6,
+    "fallbackStrategy": "cost_first",
+    "adaptiveThreshold": true
+  },
+
+  "AutoOptimization": {
+    "enabled": true,
+    "learningPeriod": 7,
+    "performanceTracking": true,
+    "costMonitoring": true,
+    "automaticTuning": false
+  },
+
+  "description": "Balanced configuration providing good mix of cost, speed, and quality",
+  "useCase": "Best for general use cases, balanced workloads, and users who want it all",
+  "optimizationTarget": "Balanced cost-performance-quality ratio",
+  "costProfile": "Medium",
+  "expectedLatency": "2-4 seconds",
+  "qualityLevel": "High",
+  "recommendedFor": [
+    "General chat",
+    "Mixed workloads",
+    "Everyday tasks",
+    "Balanced requirements"
+  ]
+}

package/templates/cost-optimized.json

@@ -0,0 +1,96 @@
+{
+  "LOG": true,
+  "LOG_LEVEL": "warn",
+  "API_TIMEOUT_MS": 30000,
+  "CUSTOM_ROUTER_PATH": "$HOME/.claude-code-router/intent-router.js",
+
+  "Providers": [
+    {
+      "name": "qwen",
+      "api_base_url": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
+      "api_key": "$QWEN_API_KEY",
+      "models": ["qwen-turbo", "qwen-plus"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 1,
+      "description": "Most cost-effective for general tasks",
+      "estimatedCost": "$0.0003 per 1K tokens"
+    },
+    {
+      "name": "gemini",
+      "api_base_url": "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
+      "api_key": "$GEMINI_API_KEY",
+      "models": ["gemini-2.5-flash"],
+      "transformer": {
+        "use": ["gemini"]
+      },
+      "priority": 2,
+      "description": "Very affordable with good quality",
+      "estimatedCost": "$0.000075 per 1K tokens"
+    },
+    {
+      "name": "glm",
+      "api_base_url": "https://api.z.ai/api/paas/v4/chat/completions",
+      "api_key": "$GLM_API_KEY",
+      "models": ["glm-4.5", "glm-4.6"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 3,
+      "description": "Cost-effective for multilingual tasks",
+      "estimatedCost": "$0.0005 per 1K tokens"
+    },
+    {
+      "name": "openai",
+      "api_base_url": "https://api.openai.com/v1/chat/completions",
+      "api_key": "$OPENAI_API_KEY",
+      "models": ["gpt-4o-mini"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 4,
+      "description": "Fallback for when quality is critical",
+      "estimatedCost": "$0.00015 per 1K tokens"
+    }
+  ],
+
+  "Router": {
+    "default": "qwen,qwen-turbo",
+    "background": "qwen,qwen-turbo",
+    "think": "gemini,gemini-2.5-flash",
+    "longContext": "gemini,gemini-2.5-flash",
+    "longContextThreshold": 100000,
+    "CostOptimization": {
+      "enabled": true,
+      "budgetAlerts": true,
+      "dailyBudget": 10.0,
+      "preferCheaper": true,
+      "costTracking": true
+    }
+  },
+
+  "BudgetManagement": {
+    "dailyLimit": 10.0,
+    "monthlyLimit": 100.0,
+    "alertThreshold": 0.8,
+    "trackByProvider": true,
+    "cheapestFirstFallback": true
+  },
+
+  "CostTracking": {
+    "enabled": true,
+    "granularTracking": true,
+    "reporting": "daily",
+    "alerts": {
+      "daily": true,
+      "threshold": 5.0
+    }
+  },
+
+  "description": "Cost-optimized configuration minimizing expenses while maintaining good quality",
+  "useCase": "Best for budget-conscious users, bulk processing, and non-critical tasks",
+  "expectedSavings": "70-90% compared to premium models",
+  "costProfile": "Very Low",
+  "qualityLevel": "Good for most tasks"
+}

package/templates/development.json

@@ -0,0 +1,104 @@
+{
+  "LOG": true,
+  "LOG_LEVEL": "debug",
+  "API_TIMEOUT_MS": 15000,
+  "CUSTOM_ROUTER_PATH": "$HOME/.claude-code-router/intent-router.js",
+
+  "Providers": [
+    {
+      "name": "openai",
+      "api_base_url": "https://api.openai.com/v1/chat/completions",
+      "api_key": "$OPENAI_API_KEY",
+      "models": ["gpt-4o-mini", "gpt-4o"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 1,
+      "description": "Excellent for coding and development tasks"
+    },
+    {
+      "name": "anthropic",
+      "api_base_url": "https://api.anthropic.com/v1/messages",
+      "api_key": "$ANTHROPIC_API_KEY",
+      "models": ["claude-3-5-sonnet-latest"],
+      "transformer": {
+        "use": ["Anthropic"]
+      },
+      "priority": 2,
+      "description": "Great for debugging and architecture"
+    },
+    {
+      "name": "qwen",
+      "api_base_url": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
+      "api_key": "$QWEN_API_KEY",
+      "models": ["qwen3-coder-plus", "qwen-plus"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 3,
+      "description": "Specialized for coding tasks"
+    },
+    {
+      "name": "gemini",
+      "api_base_url": "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
+      "api_key": "$GEMINI_API_KEY",
+      "models": ["gemini-2.5-flash"],
+      "transformer": {
+        "use": ["gemini"]
+      },
+      "priority": 4,
+      "description": "Quick for documentation and simple queries"
+    }
+  ],
+
+  "Router": {
+    "default": "openai,gpt-4o-mini",
+    "background": "gemini,gemini-2.5-flash",
+    "think": "anthropic,claude-3-5-sonnet-latest",
+    "longContext": "gemini,gemini-2.5-flash",
+    "longContextThreshold": 50000,
+    "Development": {
+      "enabled": true,
+      "preferCodingModels": true,
+      "codeContextAware": true
+    }
+  },
+
+  "DevelopmentFeatures": {
+    "codeContext": true,
+    "projectAwareness": true,
+    "debugMode": true,
+    "verboseLogging": true,
+    "testMode": false
+  },
+
+  "CodeAssistance": {
+    "languages": ["javascript", "python", "typescript", "java", "go", "rust", "cpp"],
+    "frameworks": ["react", "vue", "angular", "nodejs", "django", "flask", "spring"],
+    "specializedModels": {
+      "coding": "qwen3-coder-plus",
+      "debugging": "anthropic,claude-3-5-sonnet-latest",
+      "architecture": "anthropic,claude-sonnet-4-latest",
+      "documentation": "openai,gpt-4o"
+    }
+  },
+
+  "Testing": {
+    "mockMode": false,
+    "localTesting": true,
+    "apiValidation": true,
+    "responseCaching": true
+  },
+
+  "description": "Development-optimized configuration for coding, debugging, and software engineering tasks",
+  "useCase": "Best for developers, coding assistance, debugging, and software development workflows",
+  "optimizedFor": [
+    "Code generation",
+    "Debugging",
+    "Code review",
+    "Documentation",
+    "Architecture design"
+  ],
+  "costProfile": "Medium",
+  "expectedLatency": "2-5 seconds"
+}

package/templates/performance-optimized.json

@@ -0,0 +1,88 @@
+{
+  "LOG": true,
+  "LOG_LEVEL": "info",
+  "API_TIMEOUT_MS": 20000,
+  "CUSTOM_ROUTER_PATH": "$HOME/.claude-code-router/intent-router.js",
+
+  "Providers": [
+    {
+      "name": "gemini",
+      "api_base_url": "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
+      "api_key": "$GEMINI_API_KEY",
+      "models": ["gemini-2.5-flash"],
+      "transformer": {
+        "use": ["gemini"]
+      },
+      "priority": 1,
+      "description": "Fastest responses for time-critical tasks"
+    },
+    {
+      "name": "openai",
+      "api_base_url": "https://api.openai.com/v1/chat/completions",
+      "api_key": "$OPENAI_API_KEY",
+      "models": ["gpt-4o-mini", "gpt-4o"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 2,
+      "description": "Balance of speed and capability"
+    },
+    {
+      "name": "qwen",
+      "api_base_url": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
+      "api_key": "$QWEN_API_KEY",
+      "models": ["qwen-turbo", "qwen-plus"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 3,
+      "description": "Fast for simple tasks"
+    },
+    {
+      "name": "anthropic",
+      "api_base_url": "https://api.anthropic.com/v1/messages",
+      "api_key": "$ANTHROPIC_API_KEY",
+      "models": ["claude-3-5-haiku-latest", "claude-sonnet-4-latest"],
+      "transformer": {
+        "use": ["Anthropic"]
+      },
+      "priority": 4,
+      "description": "High quality when needed"
+    }
+  ],
+
+  "Router": {
+    "default": "gemini,gemini-2.5-flash",
+    "background": "qwen,qwen-turbo",
+    "think": "openai,gpt-4o-mini",
+    "longContext": "gemini,gemini-2.5-flash",
+    "longContextThreshold": 100000,
+    "Performance": {
+      "enabled": true,
+      "timeoutMs": 15000,
+      "retryAttempts": 2,
+      "preferLowLatency": true
+    }
+  },
+
+  "PerformanceOptimization": {
+    "enableCaching": true,
+    "cacheTTL": 300,
+    "maxConcurrentRequests": 5,
+    "compressionEnabled": true,
+    "smartTimeout": true
+  },
+
+  "HealthChecks": {
+    "enabled": true,
+    "interval": 30000,
+    "timeout": 5000,
+    "failureThreshold": 3,
+    "recoveryTimeout": 60000
+  },
+
+  "description": "Performance-optimized configuration prioritizing speed and low latency",
+  "useCase": "Best for real-time applications, chatbots, and time-sensitive tasks",
+  "expectedLatency": "500-2000ms",
+  "costProfile": "Low to Medium"
+}

package/templates/quality-focused.json

@@ -0,0 +1,105 @@
+{
+  "LOG": true,
+  "LOG_LEVEL": "debug",
+  "API_TIMEOUT_MS": 60000,
+  "CUSTOM_ROUTER_PATH": "$HOME/.claude-code-router/intent-router.js",
+
+  "Providers": [
+    {
+      "name": "anthropic",
+      "api_base_url": "https://api.anthropic.com/v1/messages",
+      "api_key": "$ANTHROPIC_API_KEY",
+      "models": ["claude-sonnet-4-latest", "claude-3-5-sonnet-latest"],
+      "transformer": {
+        "use": ["Anthropic"]
+      },
+      "priority": 1,
+      "description": "Highest reasoning and analytical capability",
+      "estimatedCost": "$15.00 per 1M input tokens"
+    },
+    {
+      "name": "openai",
+      "api_base_url": "https://api.openai.com/v1/chat/completions",
+      "api_key": "$OPENAI_API_KEY",
+      "models": ["gpt-4o", "o1", "gpt-4-turbo"],
+      "transformer": {
+        "use": []
+      },
+      "priority": 2,
+      "description": "Strong coding and problem-solving abilities",
+      "estimatedCost": "$5.00 per 1M input tokens"
+    },
+    {
+      "name": "gemini",
+      "api_base_url": "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
+      "api_key": "$GEMINI_API_KEY",
+      "models": ["gemini-2.5-pro", "gemini-2.5-flash"],
+      "transformer": {
+        "use": ["gemini"]
+      },
+      "priority": 3,
+      "description": "Good balance of quality and speed",
+      "estimatedCost": "$1.25 per 1M input tokens"
+    },
+    {
+      "name": "openrouter",
+      "api_base_url": "https://openrouter.ai/api/v1/chat/completions",
+      "api_key": "$OPENROUTER_API_KEY",
+      "models": [
+        "anthropic/claude-3.5-sonnet",
+        "openai/gpt-4-turbo",
+        "google/gemini-pro-1.5"
+      ],
+      "transformer": {
+        "use": ["openrouter"]
+      },
+      "priority": 4,
+      "description": "Access to specialized high-quality models",
+      "estimatedCost": "Variable"
+    }
+  ],
+
+  "Router": {
+    "default": "anthropic,claude-sonnet-4-latest",
+    "background": "gemini,gemini-2.5-pro",
+    "think": "openai,o1",
+    "longContext": "gemini,gemini-2.5-pro",
+    "longContextThreshold": 200000,
+    "QualityOptimization": {
+      "enabled": true,
+      "preferAccuracy": true,
+      "maxRetries": 3,
+      "temperature": 0.3,
+      "topP": 0.9
+    }
+  },
+
+  "QualitySettings": {
+    "temperature": 0.3,
+    "maxTokens": 4096,
+    "topP": 0.9,
+    "frequencyPenalty": 0.1,
+    "presencePenalty": 0.1,
+    "responseFormat": "auto"
+  },
+
+  "EnhancedReasoning": {
+    "enabled": true,
+    "chainOfThought": true,
+    "selfCritique": true,
+    "multiplePerspectives": true
+  },
+
+  "description": "Quality-focused configuration prioritizing accuracy and capabilities over cost/speed",
+  "useCase": "Best for critical applications, complex reasoning, research, and professional work",
+  "expectedQuality": "Highest available",
+  "costProfile": "High",
+  "expectedLatency": "3-10 seconds",
+  "recommendedFor": [
+    "Complex analysis",
+    "Creative writing",
+    "Code review",
+    "Strategic planning",
+    "Research tasks"
+  ]
+}