A trading journal where the numbers reconcile

The failure mode of most journals is quiet disagreement: the P&L on a trade, the win rate on a dashboard and the figure in an exported report are computed in three different places and drift apart. MetaTradee computes gross P&L, net P&L and the planned reward-to-risk ratio once, on the server, from one definition. Every screen reads that result. A realised R-multiple is not among them — reward-to-risk describes the plan, and the two are not interchangeable.

Money is stored in exact-numeric types rather than floating point, so repeated arithmetic does not accumulate the fractional-cent drift that makes a year-end total disagree with the sum of its trades.

Two routes, and they end in the same place. Log a trade by hand — entry, exit, quantity, fees, context — or import a statement you exported from your platform.

• Statements import as CSV or JSON files
• Ready-made column mappings for MetaTrader 4, MetaTrader 5, cTrader, DXtrade, Match-Trader, TradeLocker
• A generic mapping for any other statement, with columns you assign yourself
• Delimiter, locale-formatted numbers and platform date formats detected automatically
• Malformed rows are captured with their row number and the reason, never silently dropped

It is file import, not broker sync. You export the file; MetaTradee never asks for broker credentials, an investor password or an API key, because it never connects to your account.

Upload, map columns, then preview. The preview is the dry run — it validates every row against the same schema the journal uses, shows what would be written, flags duplicates, and writes nothing. You confirm, and only then does anything land. The server re-validates on commit, so a tampered client cannot bypass it.

Duplicates are caught by content hash, against both your existing trades and other rows in the same file. Re-importing an overlapping statement is safe.

A journal that only stores rows is a spreadsheet with extra steps. Trades carry tags, notes and screenshots; the calendar breaks results down by day, session and hour with timezone and DST handled; playbooks are versioned and become immutable once used, so adherence is measured against the rules that actually applied at the time.

Psychology entries record the emotions and habits around a trade and feed a transparent discipline score that rewards process rather than profit. Those entries are private by construction and are never shown to workspace admins without your explicit opt-in.

Every record is scoped to you by row-level security in the database — not by a filter in application code that a bug could skip. Reports are shareable only when you choose to share one, and report sharing is withheld during a trial so a link cannot outlive it.

Backtesting is not available — neither manual nor automated — and it is not sold as part of any plan. What ships is trade replay, which steps through real recorded sessions bar by bar. Prop-firm rule monitoring is also not built. Saying so here is cheaper for both of us than a refund request later.