📓 AcTrade — Trading Journal

multi-asset import · fill-level analytics · dark data-dense dashboard

A production-grade, TradeZella-style trading journal: import broker/exchange executions, reconstruct every trade, and see honest per-trade metrics on your edge. יומן מסחר ברמת production בסגנון TradeZella: ייבוא ביצועים מברוקר/בורסה, שחזור כל עסקה, ומדדים כנים ברמת העסקה על ה-edge שלך.

---

🌍 What is this? · מה זה?

🇬🇧 English A production-grade trading-journal SaaS. Traders import their broker/exchange execution history; AcTrade reconstructs every trade (average-cost, walk-to-flat), computes performance analytics, and presents them in a dark, data-dense dashboard. Most traders never measure their own behavior. By turning raw executions into honest, per-trade metrics, AcTrade makes the source of P&L visible and improvable. For active retail traders across crypto (Binance/Coinbase/Bybit), CFD/forex (MT4/MT5), futures (CME/CBOT/NYMEX/COMEX), and equities.

🇮🇱 עברית יומן מסחר SaaS ברמת production. סוחרים מייבאים היסטוריית ביצוע מהברוקר/בורסה; AcTrade משחזר כל עסקה (עלות ממוצעת, walk-to-flat), מחשב אנליטיקת ביצועים ומציג בדשבורד כהה וצפוף-נתונים. רוב הסוחרים לא מודדים את ההתנהגות של עצמם. בהפיכת ביצועים גולמיים למדדים כנים ברמת העסקה, AcTrade הופך את מקור ה-P&L לגלוי ולבר-שיפור. מיועד לסוחרים קמעונאיים בקריפטו (Binance/Coinbase/Bybit), CFD/פורקס (MT4/MT5), חוזים עתידיים (CME/CBOT/NYMEX/COMEX) ומניות.

---

✨ Features

Feature	What it does
🔀 Multi-asset import	Binance/Coinbase/Bybit fills · MT4/MT5 HTML + CSV · futures (point-value) · generic CSV. Content-hash dedup; malformed rows reported.
🧮 Trade reconstruction	Average-cost walk-to-flat: scale-in/out, shorts, position flips, fee proration. Exact decimal math (no float drift).
📊 Analytics	Net P&L, win rate, profit factor, avg R-multiple, max drawdown (abs + %), equity curve, timezone-aware daily P&L calendar heatmap.
📋 Trade log	Sortable, filterable (symbol/side/status), detail view with executions ladder + editable journal notes.
⚙️ Accounts & settings	Multiple accounts, base currency, profile timezone driving the calendar.
🔐 Auth	Google OAuth + email/password; Row-Level Security isolates every user's data.

---

🧱 Tech stack

Layer	Choice
Framework	Next.js 16 (App Router, Proxy not Middleware), TypeScript, React 19
Styling	Tailwind CSS v4 · hand-written shadcn-style primitives
Charts / tables	Recharts · TanStack Table · date-fns
Backend	Supabase (Postgres + Auth, RLS)
Math / parsing	decimal.js · papaparse · node-html-parser · zod
Tests	Vitest (48 tests)
Deploy	Vercel via GitHub git integration (CI/CD on push to main)

---

🚀 Getting started (local)

npm install
# .env.local already holds the Supabase URL + publishable key
npm run dev      # http://localhost:3000
npm run build    # production build / typecheck gate
npm run lint
npm test         # vitest (48 tests)

---

🗂️ Project structure

src/
  app/
    (dashboard)/            # protected: layout (auth gate) + dashboard, trades[/id],
                            #   import, accounts, settings
    auth/callback/route.ts  # OAuth code exchange
    login/                  # auth UI
    layout.tsx · globals.css (theme) · page.tsx (landing)
  proxy.ts                  # Next 16 Proxy — Supabase session refresh
  actions/                  # server actions: auth, import, accounts, trades, profile, preferences
  components/               # ui/ dashboard/ charts/ trades/ accounts/ import/ settings/ auth/
  lib/
    supabase/{client,server,session}.ts
    trades/{reconstruct,pnl}.ts
    metrics/metrics.ts
    import/{columns,fills,completed,metatrader,contracts,dedupe,index}.ts
    data/trades.ts · auth.ts · utils.ts · format.ts
  types/{database,trading}.ts

DB schema + RLS live in the Supabase project actrade (migrations 0001_init_schema, 0002_security_hardening, applied via the Supabase MCP). DB types generated into src/types/database.ts.

🏛️ Architecture notes

• Executions are the source of truth (immutable fills). Trades are derived (average-cost), with cached metric columns for fast dashboards. Re-imports dedup by content hash never double-count.
• Multi-asset from the schema up: asset_class + per-instrument multiplier; futures carry point_value / tick_size.
• RLS on every table (auth.uid() = user_id).

📌 Status & known limitations

v1 is build-complete: all five phases shipped, and npm run build, npm run lint, and the 48-test Vitest suite are green. Tracked limitations, stated rather than hidden:

• EU-locale numbers are not auto-detected (locale is per-import and explicit).
• Cross-batch incremental closes are not re-reconstructed.
• Futures contract_expiry is left null; point_value / tick_size are populated.
• The trade log is single-page (no pagination yet).
• The live deployment runs on *.vercel.app; no custom domain yet.

Current state, decisions, and the dated history live in STATUS.md, DECISIONS.md, and PROGRESS.md.

---

⚠️ This is Next.js 16 APIs differ from older versions (Middleware → Proxy, async cookies()). See AGENTS.md and node_modules/next/dist/docs/ before writing framework code.

---

Built by @www8351 · Public repo: github.com/www8351/Tradezella-Example (developed privately as acTrade)

_{Executions are truth · trades are derived · exact decimal math.}