How jdhayesbc/awareness works
Awareness provides persistent identity and memory for AI entities. The central documented workflow captures conversations, extracts and stores meaning across searchable memory layers, then restores continuity across sessions and channels. The PPS HTTP server is the runtime entry point for memory tools; its excerpt confirms raw-message capture, while the index shows its references to other layers. Haven supplies private chat, and image generation and repository RAG are distinct product capabilities. Several subsystem links are documented only at a high level; those are marked unverified rather than inferred from filenames.
Main entry points: server_http.py PPS HTTP Server · server.py Haven Server · lyra_daemon.py Lyra Daemon · main.py RAG Service
Components 20
Entity Runtime
- Lyra Daemon
daemon/lyra_daemon.py - Claude Invoker
daemon/cc_invoker/invoker.py
Persistent Memory
- PPS HTTP Server
pps/docker/server_http.py - Raw Capture
pps/layers/raw_capture.py - Core Anchors
pps/layers/core_anchors_chroma.py - Knowledge Graph
pps/layers/rich_texture_v2.py - Custom Graph
pps/layers/custom_graph.py - Rolling Summaries
pps/layers/message_summaries.py - Categorical Inventory
pps/layers/inventory.py - Technical RAG
pps/layers/tech_rag.py
Haven Chat
- Haven Server
haven/server.py - Haven Database
haven/db.py - Haven Bridge
haven/bridge.py - Haven Bot
haven/bot.py
Retrieval and Media
- Image Pipeline
image_gen/pipeline.py - Image Renderers
image_gen/renderers/base.py - RAG Service
rag-service/app/main.py - RAG Engine
rag-service/app/rag_engine.py - RAG Metadata Store
rag-service/app/storage.py
Other components
- AI Entity
Connections 19
- AI Entity → Lyra Daemon: uses sessions
- Lyra Daemon → Claude Invoker: invokes model
- Lyra Daemon → PPS HTTP Server: requests memory
- PPS HTTP Server → Raw Capture: stores messages
pps/docker/server_http.py - PPS HTTP Server → Core Anchors: uses layer
pps/docker/server_http.py - PPS HTTP Server → Knowledge Graph: uses layer
pps/docker/server_http.py - PPS HTTP Server → Rolling Summaries: uses layer
pps/docker/server_http.py - PPS HTTP Server → Categorical Inventory: uses layer
pps/docker/server_http.py - PPS HTTP Server → Technical RAG: uses layer
pps/docker/server_http.py - PPS HTTP Server → Custom Graph: uses graph
pps/docker/server_http.py - AI Entity → Haven Server: chats through
- Haven Server → Haven Database: reads and writes
haven/server.py - Haven Server → Haven Bridge: uses bridge
haven/server.py - Haven Bot → Haven Server: connects chat
- AI Entity → Image Pipeline: requests images
- Image Pipeline → Image Renderers: dispatches render
image_gen/pipeline.py - AI Entity → RAG Service: queries documents
- RAG Service → RAG Engine: runs retrieval
rag-service/app/main.py - RAG Service → RAG Metadata Store: stores metadata
rag-service/app/main.py