How susesh10/project-aether-js works
Project Aether is a personal AI companion: users chat through a React interface, and FastAPI routes messages to Akari’s Groq-backed reply service. The backend also extracts durable user facts, stores memory and tasks in MongoDB, and exposes search, task, memory-reset, and speech endpoints. The frontend has browser speech recognition and requests spoken replies. Blender assets represent a separate, early character-modeling effort; the README describes references and head-stage files, but no runtime integration. The graph follows shown calls and documented capabilities; unsampled Blender and frontend styling internals are not inferred.
Main entry points: main.py FastAPI routes · ai.py AI reply service · App.tsx Chat UI · database.py MongoDB connection
Components 18
Chat interface
- Chat UI
frontend/src/App.tsx
Conversation services
- FastAPI routes
backend/app/main.py - Chat schema
backend/app/schemas/chat.py - AI reply service
backend/app/services/ai.py - Runtime settings
backend/app/core/config.py - Search and time tools
backend/app/services/tools.py - Speech synthesis
backend/app/services/voice.py
Memory and tasks
- Fact extraction
backend/app/services/memory_extractor.py - Memory and tasks
backend/app/services/memory.py - MongoDB connection
backend/app/core/database.py
Character assets
- Character references
Blender/references - Head model files
Blender/Akari_model/head
Other components
- Companion user
- Groq LLM
- MongoDB
- SearxNG
- Edge TTS
- Browser speech recognition
Connections 20
- Companion user → Chat UI: sends messages
frontend/src/App.tsx - Chat UI → FastAPI routes: posts chat
frontend/src/App.tsx - FastAPI routes → Chat schema: validates requests
backend/app/main.py - FastAPI routes → AI reply service: requests reply
backend/app/main.py - AI reply service → Groq LLM: calls model
backend/app/services/ai.py - AI reply service → Runtime settings: resolves model
backend/app/services/ai.py - FastAPI routes → Fact extraction: extracts facts
backend/app/main.py - FastAPI routes → Memory and tasks: reads and updates
backend/app/main.py - AI reply service → Memory and tasks: uses memory and tasks
backend/app/services/ai.py - Memory and tasks → MongoDB connection: uses collection
backend/app/services/memory.py - MongoDB connection → Runtime settings: loads connection settings
backend/app/core/database.py - MongoDB connection → MongoDB: connects to
backend/app/core/database.py - AI reply service → Search and time tools: dispatches tools
backend/app/services/ai.py - Search and time tools → SearxNG: queries search
backend/app/services/tools.py - FastAPI routes → Search and time tools: serves search
backend/app/main.py - Chat UI → FastAPI routes: requests speech and reset
frontend/src/App.tsx - FastAPI routes → Speech synthesis: generates speech
backend/app/main.py - Speech synthesis → Edge TTS: synthesizes audio
backend/app/services/voice.py - Chat UI → Browser speech recognition: recognizes voice input
frontend/src/App.tsx - Character references → Head model files: informs modeling
README.md