reasoningbank-with-agentdb

Enables adaptive learning with 150x faster vector DB. Tracks trajectories, judges outcomes, distills memories, and improves decision-making.

• Adaptive learning via ReasoningBank with AgentDB
• Tracks agent execution trajectories
• Judges success of actions (verdict judgment)
• Distills experiences into high-level patterns

• 150x faster pattern retrieval
• 500x faster batch operations
• Sub-1ms memory access
• Trajectory tracking
• Verdict judgment
• Memory distillation
• Pattern recognition
• 100% backward compatibility

1. 1.Initialize AgentDB with `npx agentdb@latest init`
2. 2.Store experiences using `insertPattern` with `enableLearning: true`
3. 3.Retrieve similar experiences with `retrieveWithReasoning`
4. 4.Use PatternMatcher, ContextSynthesizer, and MemoryOptimizer modules