Charts live in one place. Execution lives in another. That split is fine until the day you need to act inside a few seconds, and the two windows disagree about what the price even is. Linking a charting platform to a broker account fixes part of that and creates new problems.
This walkthrough assumes you already have an account somewhere and a chart layout you like. What follows is the order I’d do things in, along with the fields that matter and the two mistakes that cost me an afternoon each. Most of it is reading small print before clicking connect.
Start with what the broker supports
Not every account can be traded from a chart window. Some expose full order entry. Others offer nothing beyond a data feed. The support page usually says which, in a paragraph nobody reads. Find it before you spend an evening configuring something your account type can’t do. It takes a minute to read.
The second thing to check is the instrument list. A chart symbol and a broker ticker can look identical and still refer to different contracts, with different session hours and a different tick value. I’ve placed an order on what I assumed was the same instrument and regretted it within a minute.
Reading the integration page before you click connect
Integration pages have a habit of listing what works and burying what doesn’t. Look specifically for whether stop orders rest on the broker server or inside your browser session. That single line decides whether your protection survives a closed laptop, and it’s the difference between an inconvenience and a genuinely bad morning.
The other line to find is about order types. Bracket orders that attach a stop and a target to an entry are common inside a broker terminal and less common through a chart connection. If yours doesn’t carry them across, you’ll place the protective order by hand. Fine, if you know in advance.
Linking the account without breaking anything
The connection itself is usually two clicks and a login window. The part worth slowing down for is what permissions you hand over. A key that can open and close positions is what you need. A key that can also move funds is not, and that’s worth a message to support.
Enable two-factor on the broker account before connecting, not after. Then log the date somewhere. Keys expire, sometimes quietly, and a connection that stopped working last Tuesday explains a lot of mysterious behavior once you know when it happened. A short note in the same file as your trading rules is enough.
Where the key goes and where it never should
The key belongs in the broker connection dialog and nowhere else. No chat, no screenshot, no third-party tool promising to mirror your trades. This sounds obvious until someone helpful in a community offers to check your setup for you, and the request feels reasonable because they’ve been answering questions politely for weeks.
If something looks off, revoke first and investigate after. Revocation is instant and costs you a reconnection. The alternative costs a great deal more. Treat the key list in your account settings the way you’d treat a spare set of house keys, meaning you know who holds them and how many exist.
Alerts, orders and the gap between them
An alert is a message. An order is an instruction. Confusing the two is the most common setup error I see. A price alert firing at a level doesn’t put you in a position, while an alert wired to a webhook does, which is how a test opens a real position.
Test webhooks on a practice account first, and test them during quiet hours rather than around a scheduled release. Send one alert. Watch what appears in the broker terminal. Check the symbol and the size in the order history before you assume the whole chain behaves the way the documentation says it does.
Why the fill price isn’t the alert price
When my alerts kept firing a beat late last winter, a forum thread pointed me toward this breakdown of routing delays, and the explanation was less dramatic than expected. The alert evaluates on candle close, the message travels, the broker queues it, and price keeps moving through every step of that chain.
Knowing the size of that gap matters more than eliminating it. Time a few test signals and note the difference between the alert timestamp and the fill timestamp in the order history. If the gap is wider than the move you’re chasing, chart-based execution isn’t the right tool for that setup.
Sometimes two separate windows is the right answer
There’s a version of this where you connect nothing. You read the chart on one screen and place orders by hand in the broker terminal on the other. It’s slower by a few seconds, and for anything held longer than an hour those seconds don’t matter. Plenty of people work this way.
The automation is worth having when your setup depends on reacting at a specific level while you’re away from the desk, and it becomes a liability when it lets you act faster than you can think. I’ve switched mine off for weeks at a time after noticing it mostly produced unplanned trades.
