Rajandran R Creator of OpenAlgo Sketchmaker, RagZ. Building GenAI Applications. Telecom Engineer turned Full-time Derivative Trader. Mostly Trading Nifty, Banknifty, High Liquid Stock Derivatives. Trading the Markets Since 2006 onwards. Using Market Profile and Orderflow for more than a decade. Designed and published 100+ open source trading systems on various trading tools. Now the recent curiosity on building Trading AI Agents.

AmiBroker 7.00 BETA Is Now an AI-Powered Coding Assistant

3 min read

AI chip powering a code editor beside a candlestick chart for AmiBroker AFL coding

AmiBroker 7.00 adds an AI code assistant to the AFL Editor that can write AFL from a plain-English description, explain formulas, find bugs and suggest fixes. You choose the model behind it: OpenAI, Google Gemini or a private local LLM running on your own PC. This post first covered the 7.00.0 BETA; the feature has since shipped in the official 7.00.1 release.

Key Takeaways

  • The AI Code Assistant lives inside the AFL Editor and can generate, explain, analyse and fix AFL code.
  • Cloud models (OpenAI GPT-4.1 mini, Gemini 2.5 Flash) or a local model through LM Studio can power it.
  • You accept or reject every suggestion, so you stay in control of your code.
  • AmiBroker 7.00.1 is now the official version, and the change log already lists 7.10.0.

AmiBroker 7.00.0 BETA introduces a notable shift in how traders can work with AFL (AmiBroker Formula Language), thanks to the new AI Code Assistant feature built directly into the AFL Editor. While the update still maintains the familiar structure of the platform, this version leans toward making AFL scripting more accessible, especially for those who aren’t seasoned coders.

AmiBroker 7.00 AI Code Assistant in the AFL Editor

AmiBroker 7.00 AI Code Assistant: A Practical Addition

The most significant change in this release is the integration of LLMs (Large Language Models) into the AFL coding environment. Users can connect cloud-based models such as OpenAI GPT-4.1 mini or Google’s Gemini 2.5 Flash, or even choose to run a local model like Qwen using tools such as LM Studio.

AI Bootcamp 2.0 for traders and investors

AI FOR TRADERS

AI Bootcamp 2.0 for traders and investors

Use ChatGPT, Claude and AI agents for research, coding and automation in your trading workflow.

  • Prompting and agent workflows for markets
  • Build tools without being a full-time coder
  • 6 courses in one bootcamp
See the bootcamp →by Marketcalls
AmiBroker AI Code Assistant LLM provider settings

Once configured, the Assistant appears in the AFL Editor sidebar, where it can:

  • Autocomplete code based on prompts
  • Explain AFL expressions
  • Suggest or fix syntax
  • Offer code snippets for common trading strategies

According to AmiBroker’s official “What’s new” page, the assistant writes code from your English description, explains any formula you give it, analyses your code’s strengths and weaknesses, fixes bugs and errors, and suggests improvements.

You’re given the choice to accept or reject AI suggestions before they’re applied, ensuring manual control remains central to the process. The implementation is functional and lightweight, and while not flawless, it shows promise for improving productivity, especially for users writing repetitive or boilerplate code.

AmiBroker AI Code Assistant generating AFL code

Setup Flexibility and Privacy Options

AmiBroker gives you flexibility in choosing which AI to use. You can:

  • Use OpenAI or Google Gemini APIs for better performance and higher accuracy.
  • Run an offline/local model like Qwen for privacy, assuming your machine meets GPU requirements.

Users concerned about sharing data with external servers can opt for local LLMs, though performance and quality will vary depending on the model and hardware.

AmiBroker AI Code Assistant explaining AFL code

How to Get Better AFL Code from the AI Assistant

  • Be specific: state the timeframe, entry and exit rules, stoploss, position sizing and whether you want a chart, exploration or backtest.
  • Ask for one step at a time: first the indicator, then the Buy/Sell logic, then the backtest settings.
  • Always verify: run the code, plot the signals and check a few trades by hand. LLMs can produce code that compiles but does not do what you asked.
  • Watch for look-ahead bias: check any use of Ref() with positive values, and confirm whether signals are traded on the same bar or the next bar.

I have shared similar lessons in converting SSL Hybrid from Pine Script to AFL using an OpenAI reasoning model and why AI hallucinates and how traders can catch it.

AFL and Platform Improvements (Briefly)

Beyond the AI features, AmiBroker 7.00 also includes a range of under-the-hood improvements:

  • More control in backtests via SetOption("MaxEntryRank", 10)
  • Category and tag support for symbol management
  • Extended date handling (StrToDateTime supports years up to 2037)
  • Better chart annotations and trade arrow placement
  • Several bug fixes and quality-of-life improvements in the AFL Editor and New Analysis window

The official notes also mention an integrated software updater, colour-coded symbol tags, the SetOpacity function for transparent plots, AddSummaryRows in all analysis modes and a thread-safe fappend function (source).

These changes are mostly incremental but useful, especially for power users working on more complex strategies.

Licensing and Access

AmiBroker 7.00 is a free upgrade for users who registered after July 12, 2023. Others will need to purchase a license upgrade.

Update: the official 7.00.1 release announcement states that the free upgrade applies to users who registered AmiBroker after December 22, 2023. Check the Account Information page on the AmiBroker members site to confirm your eligibility.

To use the AI Assistant beyond 5 completions per session, you also need a valid AFL Code Wizard license. Without it, AI functionality is limited.

Cloud LLM usage may incur token-based costs (depending on the provider), though Gemini currently offers a generous free tier.

From 7.00 BETA to 7.00.1 and 7.10

The 7.00.0 BETA is no longer supported; AmiBroker now recommends the official 7.00.1 release, which fixes StrToDateTime(“”) to return Null again and improves the Batch Scheduler. The AmiBroker release notes already lists version 7.10.0, with additions such as custom N-month intervals and a warning for slow chart formulas.

Final Thoughts

AmiBroker 7.00 doesn’t reinvent the platform, but it does add a meaningful layer of support for users who want to work faster, or learn AFL as they go. The AI Assistant is not perfect, but it’s a useful companion and a step forward for those who spend a lot of time in the AFL editor.

If you’re already comfortable with AFL, this may not drastically change your workflow. But if you’re someone who appreciates coding assistance, quick fixes, or explanation on-the-fly, this new feature is worth exploring.

Frequently Asked Questions

Which AI models work with the AmiBroker AFL Code Assistant?

You can connect OpenAI models such as GPT-4.1 mini, Google Gemini models such as Gemini 2.5 Flash, or a local model such as Qwen running through LM Studio.

Is the AmiBroker 7 upgrade free?

It is free for users whose registration falls within the eligibility window stated by AmiBroker; others need to buy an upgrade. Check your Account Information page on the AmiBroker members site.

Can I use the AmiBroker AI assistant offline?

Yes. Running a local LLM keeps your code on your own computer, but you need suitable hardware and the code quality depends on the model.

Does the AI assistant replace learning AFL?

No. It speeds up writing and debugging, but you still need to understand AFL well enough to check that the generated code and backtest logic are correct.

Python Elite membership

MEMBERSHIP

Python Elite membership

15 Python, AI and algo trading programs in one plan, including QuantFlow and AI Bootcamp live mentorship.

  • Python from fundamentals to trading apps
  • Backtesting, broker APIs and OpenAlgo
  • Live QuantFlow and AI Bootcamp sessions
View Python Elite →by Marketcalls

Learn by watching: free videos every week

Market Profile, order flow, Python, AI and OpenAlgo tutorials on YouTube.

Rajandran R Creator of OpenAlgo Sketchmaker, RagZ. Building GenAI Applications. Telecom Engineer turned Full-time Derivative Trader. Mostly Trading Nifty, Banknifty, High Liquid Stock Derivatives. Trading the Markets Since 2006 onwards. Using Market Profile and Orderflow for more than a decade. Designed and published 100+ open source trading systems on various trading tools. Now the recent curiosity on building Trading AI Agents.

2 Replies to “AmiBroker 7.00 BETA Is Now an AI-Powered Coding Assistant”

  1. Thanks for the article Rajandran sir. Does Amibroker work with Macbook M4 using parallels? Will the RT datafeed like GFDL work flawlessly or are there any catches?

Leave a Reply

Get Notifications, Alerts on Market Updates, Trading Tools, Automation & More