Inside the PraxAI Dashboard: What the Walkthrough Video Actually Shows

The first screen in the video does not ask what you trade. It asks who evaluates you. You pick your prop firm before you pick anything else, and the two minutes that follow are the consequence of that one choice: the files that firm selection produces, then the robot running under those numbers on a XAUUSD chart inside MetaTrader.

There is no presenter and no studio footage. It was recorded on the September 2026 build of the product, which matters because the dashboard and the robots are updated most weeks. A screen recording of trading software is a snapshot, not a feature list, and anyone showing you one should say so.

Everything below is on the screen in the video, in the order it appears.

The first screen asks which prop firm you trade with. In the recording the list includes FTMO, FundedNext, FundingPips, The 5%ers, E8 and Alpha Capital Group. Appearing in a selector is not a statement about what any of them permit: whether a firm allows automated execution, and on which account types, is written in that firm's own terms on the day you read them. That order of operations is the whole design argument of the product: the firm you are evaluated by determines the limits your automation has to live inside, so the firm is chosen before anything else is configured.

Most trading software asks about your strategy first and treats risk as a setting you adjust afterwards. In a prop firm evaluation that is backwards. A strategy that is profitable but breaches a daily loss limit on one bad session has failed the only test that was being run.

After the firm is selected, the dashboard hands over the files for that firm instead of opening a blank configuration screen. The video shows this taking about a minute, which is realistic for the download and attach steps and does not include opening a broker account or setting up a server.

If you want the longer explanation of why firm selection drives everything downstream, it is written out in the prop firm challenge checklist for automated trading.

A setfile is a text file of a few kilobytes that tells an expert advisor where its walls are: the percentage it may risk per position, the daily loss figure at which it stops opening anything, the minutes around a scheduled release when it does not trade at all. You can open one in Notepad and read every number in it, which is the reason it is worth reading. In the video the setfile arrives already filled in for the firm that was selected on the previous screen. The values in that file are our own reading of that firm's published limits at the time the file was built, not a statement issued by the firm, so check them against your own rulebook before the first trade.

This is the least cinematic moment in the recording and the most important one. Two traders can run the same robot on the same chart and get opposite outcomes, because one of them is running numbers that sit under their firm's thresholds and the other never changed the default.

It is also the part of the product that has to be maintained. Prop firms revise their rules, and a setfile that was correct in March is not automatically correct in September. That maintenance work is invisible in a video, which is a fair reason to ask a vendor how their files get updated and how you are told when they change.

What a setfile cannot do is change what your firm allows. Whether a given firm permits expert advisors at all, and under what conditions, is written in that firm's own terms, and it is the one thing you should read yourself instead of taking it from any video, including ours.

The final section of the recording is a MetaTrader window with the robot attached to a XAUUSD chart. Gold is the instrument shown because it is the least forgiving thing in the video. XAUUSD can travel more in one London hour than many pairs travel in a session, which means a loose daily limit shows itself in minutes instead of weeks.

In the September build the work is split between two engines. The PASS Engine is the component that runs during the evaluation phase, pursuing the profit target while holding daily loss and total drawdown under the thresholds in the setfile. The PAYOUT Engine takes over once the account is funded, when the objective changes from hitting a target to keeping an account alive long enough to reach a payout cycle. Newer builds break the same work into narrower components, which is described in the three stage overview. The recording shows the two engine split as it stood in September 2026.

Underneath both of them sits PraxAI GUARD, and it is worth being precise about what it is. PraxAI GUARD is hard coded logic, not a model forming an opinion. It holds the firm's numbers and it stops the robot when a threshold is approached. It does not forecast, it does not weigh probability, and it cannot be talked out of a stop by a good looking setup. That is the entire point of building it that way.

The distinction matters when you compare products. Ask a vendor what halts the robot. If the answer contains the word intelligent, the answer is a model, and a model can be wrong at the worst possible moment. If the answer is a number and a line of code, you can test it on a demo account this afternoon.

Two minutes is enough to show a workflow and not enough to show an outcome, so it is worth naming the gaps here instead of leaving you to find them later.

It does not show a full evaluation from the first trade to the pass. That typically takes weeks and no screen recording compresses it honestly. It does not show a losing stretch, which every automated system has. It does not show the account being funded or a payout arriving, because those are events that belong to the trader and the firm, not to the software.

It also does not show setup friction: opening the account with the firm, installing MetaTrader, running the robot on a server so it stays on when your computer is not. The video says about a minute to connect, and that is accurate for attaching the robot, not for the whole journey from purchase to first trade.

None of that makes the recording dishonest. It makes it a product walkthrough and not a results claim, and the difference between those two is the single most useful thing to look for when you are comparing EA vendors.

If you are actually deciding whether to buy a prop firm robot, watch the recording with three questions in your hand.

First, does the software know the rules of your specific firm, or does it expect you to translate them into settings yourself? Second, what stops it when a rule is about to break, and is that mechanism a fixed rule or a prediction? Third, how does the product stay correct when your firm changes a limit next quarter?

Those three questions separate almost every product on sale, and they can all be answered from a demo. Whether a given robot makes money cannot be answered from a demo, which is exactly why so many demos are edited to look like they answer it.

Those three questions are the ones we built the product to answer, and a video is the wrong place to check our answer. The dashboard is. Our refund window is seven calendar days from the purchase date and it is conditional rather than no questions asked, with every condition published on the refund policy page, which is worth reading before a purchase instead of after one. If you would rather keep reading first, the short version of the argument is in the thirty second explainer and the failure mechanism is in why most traders fail prop firm challenges.