pil-memory-bus

What problem does it solve?

Unified memory bus for the Proactive Intelligence Layer, delivering a four-tier memory hierarchy with namespace isolation, TTL-driven storage, and vector search to preserve cross-session context, enable efficient memory consolidation, and support persistent intelligence.

Core Features & Use Cases

• Four-tier memory hierarchy: Global, Project, Session, and Task memories with TTL guidance and persistence guarantees.
• Namespace-based organization: client, pipeline, patterns, and coordination scopes to isolate and retrieve relevant data.
• Memory types and lifecycle: knowledge, context, task, and result data types with appropriate TTLs and retention policies.
• Semantic search and retrieval: vector-search-backed querying to surface related memories and patterns for decision support.
• Use cases: maintain continuity across conversations, store deal history, transcripts, and campaigns for rapid retrieval and reasoning.

Quick Start

Run your AI agent with the PIL memory bus enabled to persist cross-session context and retrieve memories using the client, deals, and meetings namespaces.