Smart-AI-Bridge

# Changelog - Smart AI Bridge v1.1.0 ## [1.1.0] - 2025-10-16 ### Added #### LocalServiceDetector - Auto-discovery of local AI services (vLLM, LM Studio, Ollama, Text-Generation-WebUI) - WSL IP detection with 3 fallback strategies - 5-minute intelligent caching with manual invalidation - Parallel endpoint testing (<6 seconds for 28 candidates) - New MCP tool: `discover_local_services` #### ConversationThreading - Multi-turn conversation management with thread IDs - Thread creation, continuation, and resume capabilities - Full-text search across conversations - Conversation analytics and history tracking - New MCP tool: `manage_conversation` #### UsageAnalytics - Session statistics and backend performance metrics - Cost analysis and optimization recommendations - Historical analytics with time ranges (1h, 24h, 7d, 30d) - JSON and Markdown export formats - New MCP tool: `get_analytics` #### Dashboard Server (Optional) - Real-time web-based monitoring interface - WebSocket updates for live metrics - Backend management capabilities - Default: DISABLED (opt-in via ENABLE_DASHBOARD=true) ### Changed - Main entry point updated to `smart-ai-bridge-v1.1.0.js` - Package description updated to include new features - Added version-specific npm scripts ### Environment Variables (New) ```bash LOCAL_AI_DISCOVERY_ENABLED=true # Default: true LOCAL_AI_DISCOVERY_CACHE_TTL=300000 # 5 minutes ENABLE_DASHBOARD=false # Default: false DASHBOARD_PORT=3456 # Default: 3456 ``` ### Backward Compatibility - 100% compatible with v1.0.0 - All existing tools and APIs unchanged - Legacy v1.0.0 server available via `npm run start:v1.0.0` --- ## [1.0.0] - 2025-09-30 Initial production release of Smart AI Bridge with multi-backend orchestration, intelligent routing, and fuzzy matching capabilities.