AmiBroker on Mac or Linux: your options, and what each one costs

By John, builder of Wise Apple · Updated August 27, 2026

You searched "AmiBroker Mac" and got a forum thread from 2011 saying to use Wine, a newer one saying Parallels, and nothing that told you which applies to you. That depends on one question, and it is not about software.

AmiBroker is a native Windows application with no Mac or Linux build, and it is one of the fastest backtesting engines an individual can buy — its array-based AFL language will scan and test across large universes faster than most things built since. If you are here, you either already use it and changed machines, or you were shopping for a backtester and hit the Windows wall. Those are different situations and they get different answers.

Option 1: Wine

Wine runs Windows programs on Mac and Linux by translating Windows system calls, with no Windows licence. People report running AmiBroker under Wine for basic charting and backtesting, and those reports are worth taking seriously — for the core engine, it often works. The breakage tends to be at the edges: data plugins that depend on Windows components, anything using OLE or ActiveX automation, printing, and occasionally rendering. Each AmiBroker update and each Wine update can change which edges break. On Apple Silicon Macs there is an additional translation layer, and the reports get thinner.

Cost: free, and an unknown number of evenings. Suits: someone who mainly needs AFL backtesting with imported data and is comfortable troubleshooting.

Option 2: a virtual machine or Parallels

Run real Windows in a window. On Intel Macs and Linux, VirtualBox or VMware with a Windows licence. On Apple Silicon, Parallels with Windows on ARM, where AmiBroker runs under Windows' own x86 emulation — one more layer, and again worth verifying for your specific setup rather than assuming parity. This is the option that works most reliably, because AmiBroker is running on the operating system it was built for. The cost is a Windows licence, a virtualisation licence on Mac, memory, and a workflow that lives in a second desktop.

Cost: licences, RAM, and some friction. Suits: the existing AmiBroker user who wants everything to just keep working.

Option 3: something that already runs where you are

If you were shopping rather than migrating, the Windows wall is a reason to look at tools built for Mac and Linux — or built for the browser, which is all of them. The Mac and Linux survey covers the category. The rest of this page is about one such tool, and it is honest about what you give up.

Side by side

Stay with AmiBroker if

• You have a library of AFL and a process built around it.
• You scan and rank across large universes; nothing here matches that.
• Raw speed is the feature you bought it for.
• You use broker plugins for execution.
• You are fine running Windows in a window, which is a perfectly good answer.

Look elsewhere if

1. You never learned AFL and do not want to.
2. You are on Apple Silicon and the translation stack makes you nervous.
3. You want machine learning as a bench of validated models rather than something you wire up.
4. You want the honest-backtest settings — no look-ahead, walk-forward, real costs — as defaults rather than as discipline.
5. Your strategies are single-instrument, and universe scanning is not what you need.

Data is where Wine breaks

The consistent theme in Wine reports is that AmiBroker's engine runs and its data plugins do not. Real-time and broker-feed plugins depend on Windows components — COM, ActiveX, vendor DLLs — that Wine translates unevenly. The practical workaround is to feed AmiBroker from files: export or download candles as CSV, import them through AmiBroker's own importer, and treat the installation as a backtester rather than a live terminal. That works for end-of-day and downloaded intraday data, and the data guide covers where to get files and what to check in them. If your workflow depends on a live plugin, Wine is the wrong route and the VM is the right one.

What carries over if you switch, and what does not

AFL does not translate into anything else, so a library of scripts is a library that stays with AmiBroker. What does carry over is the logic. A typical AFL strategy — a moving-average condition, an ADX filter, an ATR stop, a fixed holding period — is a set of decisions that any honest tool expresses differently: the condition becomes an entry rule, the filter becomes a filter, the stop becomes an exit, the period becomes a time exit. The parts that do not carry over are the ones that only AFL does well — custom backtester hooks, rotational ranking across a universe, and the raw speed that makes a thousand-symbol scan practical. If your strategies are mostly the first kind, switching is a translation. If they are mostly the second, it is a rewrite, and the VM is cheaper.

What you are actually deciding

Between continuity and a fresh start. AmiBroker under virtualisation is a slightly awkward version of a tool you already know, with all its speed intact. A browser-based lab is a different tool with a different ceiling — lower on universe scanning and raw throughput, higher on built-in validation and machine learning — and no migration path for your AFL. Neither is wrong. The mistake is choosing the second because the first felt inconvenient, and then missing the scans.

Where Wise Apple sits

Wise Apple runs in the browser on any desktop operating system — Windows, Mac, Linux, ChromeOS — with nothing to install, and native builds for all platforms are planned. Training and backtesting happen on your own machine via WebAssembly, so the "your computer does the work" property AmiBroker users value is preserved. Strategies are built visually in PowerCore Studio; WiseApple Script, a Pine-like language, covers custom indicators. Fills after the decision bar, Embargo Bars, the Walk-Forward Window and the cost models are on by default, and nineteen models across six families sit on the bench. It tests one instrument at a time, places no trades, and is early software from one builder — which is the part to weigh against two decades of AmiBroker.