Diary — Life Tracking System

"Time is your most valuable asset (till immortality)"

A personal life tracking system built on three pillars:

1. Diary — Continuous data recording. Activities → points → streaks → trends.
2. Harada Method — Top-down goal structure. Every activity maps to a life domain.
3. AI Accountability — An active partner that parses natural language input, reminds, and prevents lapse.

Architecture

See docs/ARCHITECTURE.md for the full system design.

Key principle: LLM handles natural language input parsing ONLY. Everything else is deterministic TypeScript code.

Voice/Text input (natural language)
↓
LLM: parse → structured JSON entry ← ONLY non-deterministic step
↓
TypeScript: append to data/YYYY-MM-DD.json
TypeScript: recalculate streaks, points, totals
TypeScript: regenerate dashboard.md
↓
Static HTTP server → tracker.deardiary.network

Quick Start

npm install
npm run build
npm run serve # Start dashboard server
npm run render # Regenerate dashboard.md from JSON data
npm run import # Import legacy diary app data

Project Structure

diary/
├── src/
│ ├── schema/ # TypeScript types & interfaces
│ ├── core/ # Business logic (store, streaks, points, entries)
│ ├── render/ # JSON ↔ Markdown bidirectional conversion
│ ├── server/ # Static HTTP server for dashboard
│ └── cli/ # CLI entry points
├── data/ # JSON data files (gitignored, local only)
├── docs/ # Architecture, schema, decisions, roadmap
└── tests/ # Unit tests

Tech Stack

• Language: TypeScript (strict mode)
• Runtime: Node.js 22
• Data: File-based JSON (one file per day)
• View: Markdown (generated from JSON)
• Serving: Minimal HTTP server via Tailscale Funnel
• Testing: Vitest

Phases

• Phase 0 — "Just Talk to Me": Voice/text → LLM parse → JSON → Markdown → Serve
• Phase 1 — Dashboard + Harada integration + charts
• Phase 2 — Calendar correlation, proactive AI, shareable summaries

---

Started 2026-02-11. Rebuilt from scratch on the boyangwang/diary repo.