How xargonwan/synthetic_heart works
Synthetic Heart is a modular, persistent AI-persona framework: people interact through chat interfaces or its WebUI, while the system routes turns through response generation and action handling. Its wider product includes switchable model engines, durable memory and personality, plugins for autonomous growth and integrations, and a VRM avatar experience. The graph combines README-documented capabilities with relationships supported by the supplied source index; unsampled subsystems are included where the README establishes their role, without asserting unverified internal calls.
Main entry points: interface Chat interfaces · auto_response.py Response orchestration · message_queue.py Message routing · core_initializer.py Component discovery
Components 22
Runtime and routing
- Runtime startup
main.py - Component discovery
core/core_initializer.py - Message routing
core/message_queue.py
Conversation and cognition
- Response orchestration
core/auto_response.py - Prompt assembly
core/prompt_engine.py - Action parsing
core/action_parser.py
Persistent identity
- Memory recall
core/history_engine.py - Chat history cache
core/chat_history_cache.py - Persona state
core/persona_manager.py - Persistent database
core/db.py - Emotion subsystem
plugins/emotion_manager/emotion_manager.py
Engines and plugins
- Model engine registry
core/cortex_registry.py - Model engines
engines/external_engines - External endpoints
core/external_endpoints/registry.py - Plugins and growth
plugins - Voice engines
plugins/auris_engines - Web search
plugins/web_search/search_engine.py
Interfaces and avatar
- Chat interfaces
interface - WebUI
core/webui.py - VRM avatar
frontend/src/composables/vrm
Other components
- Person
- Model provider
Connections 22
- Person → Chat interfaces: starts conversation
README.md - Chat interfaces → Message routing: submits turns
- Person → WebUI: configures Synth
README.md - Runtime startup → Component discovery: initializes components
main.py - Runtime startup → Message routing: starts processing
main.py - Component discovery → Chat interfaces: discovers interfaces
core/core_initializer.py - Component discovery → Plugins and growth: discovers plugins
core/core_initializer.py - Message routing → Response orchestration: dispatches turns
core/auto_response.py - Response orchestration → Prompt assembly: requests prompt
core/auto_response.py - Response orchestration → Action parsing: parses actions
core/auto_response.py - Response orchestration → Message routing: enqueues messages
core/auto_response.py - Prompt assembly → Memory recall: recalls context
core/prompt_engine.py - Memory recall → Chat history cache: reads history
core/history_engine.py - Prompt assembly → Model engine registry: selects engine
core/prompt_engine.py - Model engine registry → Model engines: loads engine
core/cortex_registry.py - External endpoints → Model provider: connects providers
- Persona state → Persistent database: persists state
- Plugins and growth → Persistent database: stores memories
plugins/bio_manager/bio_manager.py - Emotion subsystem → Persona state: shapes behavior
- Voice engines → Chat interfaces: enables voice
- Chat interfaces → VRM avatar: presents avatar
- Plugins and growth → Web search: offers search
plugins/rift_vessel/knowledge_client.py