Hyperformula 3.2.0 brings improvements in two important areas: Excel function coverage and large-scale computation. Those changes are particularly useful for teams using HyperFormula as a calculation engine in financial systems, analytics platforms, or AI-powered applications that rely on Excel-compatible formula semantics.

In this release, we expanded function coverage with three new functions: IRR, VALUE, and N to better support Excel-based projects, especially those involving financial calculations and explicit type conversion. At the same time, we removed a structural limitation in the dependency graph, enabling stable processing of spreadsheets with even tens of millions of cells.

IRR Function: Financial Modeling Support

Expanding financial function coverage is an important step toward making HyperFormula an even more reliable Excel-compatible engine for production use.

Widely used in financial analysis and capital budgeting, the new IRR function allows you to calculate the discount rate at which a series of cash flows results in a net present value of zero. IRR (Internal Rate of Return) has been one of the most requested financial functions, especially from teams building advanced financial applications and AI-driven simulations that rely on accurate Excel behavior. 

With this functionality, you can eliminate the need to maintain custom financial logic when comparing investments/projects on a percentage return basis.

VALUE Function: Reliable Text-to-Number Conversion

HyperFormula is often used in environments where spreadsheet inputs come from APIs, CSV files, or user input, where data isn’t always perfectly formatted. In these cases, numeric values are often stored as text, later leading to inconsistent formula results. 

With the new VALUE function, which converts a text string representing a number into a real, numeric value the engine can calculate with, you can skip the time-consuming manual data cleaning. No matter if you’re importing currency or date formats, you can expect clean and usable data for formulas, dashboards, or charts. 

N Function: Excel-Consistent Type Conversion

To provide Excel-consistent behavior, HyperFormula must replicate not only functions but also Excel’s type conversion rules. The new N function plays a small but important role in achieving this by converting a value into a number:

• Numbers remain unchanged
• Dates convert to serial numbers
• TRUE becomes 1, FALSE becomes 0
• Text becomes 0

Let’s take a look at a simple example where you need to add inline comments to your formulas without breaking the logic. 

Because N(”text”) returns to 0, you can write:

 =A1 * 1.2 + N("20% markup")

The text doesn’t affect the calculation, since "20% markup" evaluates to 0.

Adding N function strengthens compatibility in models that depend on explicit type coercion, making sure that advanced spreadsheets behave as expected when evaluated in automated systems, backend services, and AI-powered tools.

Removing Structural Limits in Large-Scale Spreadsheets

For some teams using HyperFormula, spreadsheets aren’t just large; they’re massive. We’re talking about millions of cells connected through complex dependency graphs. As usage grows, we need to make sure that the engine scales predictably under extreme workloads.

In 3.2.0, we improved the architecture of the dependency graph by removing reliance on JavaScript’s Map size limits. The internal data structures were redesigned to ensure that graph scalability is no longer constrained by that boundary.

To validate these changes in practice, we tested HyperFormula on the NYC Tree Data set, which contains over 30 million cells, and confirmed stable operation at that scale.

Special thanks to @Krizz for helping discover, report, and implement this improvement.

Supporting AI-Driven Applications

Many AI assistants and agents generate or evaluate spreadsheet formulas programmatically. In these scenarios, consistent Excel semantics and predictable execution are essential.

But the use case goes beyond spreadsheets. Some AI systems, including agent-based solutions, use Excel formula syntax to define calculations in a structured way. Instead of relying on a language model to compute results, they delegate execution to a deterministic engine to guarantee correctness and eliminate hallucinated outputs.

With expanded function coverage and improved scalability, HyperFormula can serve not only as a spreadsheet engine but also as a reliable math layer for AI agents and other AI-based systems that require accurate and reproducible results.

If you’re building AI-powered tools that rely on spreadsheet logic or hallucination-free calculation results, feel free to email us at: hello.hyperformula@handsontable.com. We’d be happy to hear about your use case and discuss how HyperFormula can support it.

Release Notes

Added

• Added a new function: IRR. #1591
• Added a new function: N. #1585
• Added a new function: VALUE. #1592

Fixed

• Fixed Error Map maximum size exceeded error when loading big spreadsheets. #1602