How williamkoller/chronos-ai works
Project Type and Purpose
• Chronos AI is a full-stack productivity assistant combining:
• A Streamlit dashboard (frontend)
• A FastAPI service (backend/API)
• A hybrid AI engine (dev-mode rules + LocalAI in production)
• A Notion integration and a lightweight local database
• Purpose: Automate and optimize personal task scheduling via intelligent algorithms and local LLM inference.
File Structure
• Top-level directories:
• api/ → FastAPI endpoints (main.py)
• dashboard/ → Streamlit UI (app.py)
• core/ → Business logic (models.py, scheduler.py)
• integrations/ → External service clients (Notion, AI dev client)
• learning/ → Feedback-based ML components (pattern_analyzer, feedback_processor)
• localai/ → LocalAI model and config files (production LLM)
• Config and orchestration:
• docker-compose.yml, Dockerfile.api, Dockerfile.dashboard for containerization
• Makefile for automation (setup, up, down, logs, tests)
• .env.example for secrets and environment variables
• chronos_knowledge.db for local SQLite storage
README for Additional Insights
• Architecture section already shows three main blocks: Dashboard ↔ API ↔ Hybrid AI Engine
• Database is PostgreSQL in production (SQLite local for dev)
• LocalAI runs on port 8080; API on 8000; Dashboard on 8501
• Hybrid AI Engine toggles between rule-based dev mode and LocalAI container based on AI_DEV_MODE
• Integrations: Notion API for task syncing
Elements
• Labels: component names, ports, key directories
• UI → API → AI Engine and Database
• API ↔ Notion API
• learning modules ← API/core → scheduler
• Rectangles for services (API, Dashboard, LocalAI)
• Clouds for external services (Notion API)