BAS Free vs Premium: What Actually Limits You and What a Project Really Costs

The “free or paid” question gets asked before a project starts, but it is better answered after the first prototype. The reason is simple: the free version is enough to build a bot, and people usually hit the limits when they start to scale one.

One note up front: specific prices and tier contents change, so they are deliberately absent here - check the developer’s site for current numbers. What is more useful is understanding what is limited and when it starts to hurt.

What the free version covers

More than people expect. The whole development environment, the action tree, modules, browser interaction, the HTTP client, the database, compilation into an application - that is enough to build and ship a real production bot.

The practical takeaway: do not buy anything before you have a prototype. Until the bot works you do not know which resource will turn out to be scarce - and very often it is not the licence tier.

Where the real limits begin

The limit people actually hit is the number of concurrent threads. That is the number premium is bought for.

But do not confuse cause with symptom. Before declaring the thread limit your bottleneck, rule out two cheaper explanations:

• The iteration is not optimised. If the bot spends two thirds of its time waiting, you are paying for copies of an idle process. Halving iteration time is equivalent to doubling threads - and costs nothing. See memory and performance.
• The machine cannot take it anyway. With 8 GB of RAM, lifting the thread cap changes nothing: you hit hardware first. Upgrading the tier while short on memory just relocates the bottleneck.

The rule: an upgrade makes sense when the thread limit would still constrain an optimised bot on an adequate machine.

Size it from the task, not the feature list

It is far more useful to ask “how much work must fit in my window” than “how many threads can I have”.

The arithmetic is easy: take the time of one iteration, the volume of work, and the deadline. A thousand accounts at 40 seconds each is 11 hours of single-threaded work. If your window is overnight, a modest thread count suffices. If the same job must finish in an hour, you need an order of magnitude more, and the tier genuinely becomes the constraint.

That calculation takes five minutes and settles the question better than any review.

The project cost is not just BAS

The most common planning mistake is counting only the software. In a real bot, the recurring spend usually sits around BAS:

• Proxies - normally the largest line in multithreaded projects, especially with mobile or residential addresses. Scales with thread count.
• Captcha solving - paid per solve. Do the volume maths: fractions of a cent each become a real number across tens of thousands of iterations.
• Email and SMS code reception - numbers and mailboxes cost money, and SMS is usually the most expensive item per account.
• A VPS - a machine that never sleeps, if the bot runs on a schedule.

On many projects these add up to noticeably more than the software itself. Comparing “free versus premium” in isolation from the infrastructure means measuring the wrong quantity.

How to decide

The order that saves money:

1. Build a prototype on the free version and confirm the task is solvable at all.
2. Time one iteration and calculate how many threads your window requires.
3. Optimise the iteration - the required thread count often halves right there.
4. Check whether the machine can host that number.
5. Only now look at whether the tier limit is what stops you.

On most projects I have seen, step five arrives later than the client expected - and almost never arrives while the bot still does not work.