Multi-broker dashboard tools you actually use

[LondonScalper]

I currently have money at three brokers and a prop evaluation running, which means four logins, four platforms and, until recently, four separate spreadsheets that I updated by hand on Sunday nights.

That routine took about 50 minutes a week and was full of copy errors. Last month I found a GBPUSD ticket counted twice and another missing completely.

What I've tried so far:

A trade copier dashboard that also shows open positions across accounts. Useful for seeing total exposure at a glance, which matters when I'm long EURUSD on one account and have forgotten a small EURGBP position on another. Less useful for history.

Exporting statements as HTML from each MT4 and parsing them with a short Python script into one CSV. This is what I use now. It runs in under a minute and gives me one file with broker, symbol, open and close times in London time, size, spread cost, commission and swap.

A paid journal service that connects to several brokers. It was nice, but it was slow to sync and I didn't like giving read access to that many accounts.

The thing I still haven't solved is a live view of combined daily P&L in one place without opening four terminals. If anyone has something that does that without asking for trading passwords, I would like to hear what you run and how reliable it has been.

[Shadow Trader]

The HTML export plus Python route is what I ended up with too. One thing I'd add to the script is a check that counts tickets per account and compares it with the number in the statement header. That would have caught your double GBPUSD and the missing one straight away.

For the live combined P&L, the cleanest option I've found without passwords is the investor login. Most MT4 and MT5 brokers give a read only password. Run one terminal per account logged in with it, have a small script in each write equity to a shared file every few seconds, and let one sheet add them up.

It isn't elegant, and it means four terminals running on the VPS, but none of them can place a trade.

The other thing I'd watch is time zones. Brokers often use different server times, so your "daily" P&L can be cut at 22:00 on one and 00:00 on another. I convert everything to London time before summing.

Does your prop platform give you an export at all, or is that the one you type in by hand?

[Fairman]

Exporting statements as HTML from each MT4 and parsing them with a short Python script into one CSV.

If any of the accounts are on MT5, there's a shortcut that skips the HTML step. The MetaTrader5 Python package can connect to a running terminal and pull history directly with history_deals_get for a date range, already as structured records with commission, swap and profit per deal. Run it once per terminal and append to the same CSV. MT4 doesn't have that package, so the statement parse is still the way there, or a small script inside MT4 that writes closed trades to a file. Either way, the step that tends to go wrong is the time conversion. Server times are usually GMT+2 or GMT+3, the server switches on US dates and London on UK dates, so a fixed offset is wrong for a few weeks a year. This year that includes 26 to 30 October.

[Shadow Trader]

Fairman, the MetaTrader5 package is a good shortcut, and there's one quirk worth knowing before trusting it. The time on each deal comes back as Unix seconds, but it's the trade server's clock, not true UTC.

So the offset problem you describe is still there, just hidden inside a number that looks like UTC. I got caught by that and had a week where every trade sat two hours early.

Rather than compute the offset, I now store it per date in a small table, taken from the last tick time on each server against my own clock. That handles the 26 to 30 October gap without any DST logic in the script.

The other difference from MT4 is that one position can become several deals. Partial closes and the in and out legs mean a raw deal count won't match a ticket count, so I group by position_id before running the count check I mentioned.

The package also connects to one terminal per Python process. With four accounts I run four short processes in turn, because switching terminals inside one script was flaky for me.

Has anyone checked whether an investor login gives the full deal history through the package? I haven't tested it, and it would decide whether the read only route covers history too.