JSON Logic Engine

Fast, Powerful, and Persistable Logic

Have you ever needed the ability to write a custom set of logic or set of rules for a particular customer? Or needed to be able to configure a piece of logic on the fly?

JSON Logic might be your solution! Designed with a lisp-like syntax, JSON Logic makes it easy to write safe instructions that can be persisted into a database, and shared between the front-end and back-end.

Check out our Documentation Here.

---

Why json-logic-engine?

This library was built as a more modern, performant, and feature-rich alternative to json-logic-js. Here's why you might choose it:

Performance

• Optimized Interpreter: Uses closures to achieve ~5x faster evaluation on average
• Logic Compilation: Compile your rules for 12.5-20x performance improvements in hot paths
• Deterministic Evaluation: Prevents redundant re-computation of deterministic logic

Features

• Async Support: First-class support for asynchronous logic evaluation
• Custom Control Structures: Define your own operators with lazy or eager evaluation semantics
• Scope Traversal: Handlebars-style ../../ traversal for accessing outer context in iterators
• Error Handling: Try/throw support with stricter error handling
• Extended Proposals: Implements additional JSON Logic specification proposals

Modern JavaScript

• Proper ESM & CJS Support: Dual-build package with proper ES Modules and CommonJS support
• Unlike json-logic-js which uses a non-standard module format (neither proper CJS nor UMD), this library works seamlessly with modern bundlers and Node.js

Security

• Better safeguards when evaluating rules from untrusted sources

Bundle Size Concerns?

If bundle size is critical for your use case, check out json-logic-engine-slim (~4kB gzipped) which provides core functionality with a smaller footprint.

---

The engine supports both synchronous & asynchronous operations, and can use function compilation to keep your logic performant at scale.

Examples:

The premise is the logic engine traverses the document you pass in, and each "object" is interpreted as an instruction for the engine to run.

logic.run({
    '+': [1,2,3,4,5]
}) // 15

If you wanted to start factoring variables, you can pass a data object into it, and reference them using the "var" instruction:

logic.run({
    '+': [11, { var: 'a' }]
}, {
    'a': 17
}) // 28

The engine will also allow you to reference variables that are several layers deep:

logic.run({
    '+': [{ var: 'a.b.c' }, 5]
}, {
    a: { b: { c: 7 } }
}) // 12

Let's explore some slightly more complex logic:

logic.run({
    'reduce': [{ var: 'x' }, { '+': [{ var: 'current' }, { var: 'accumulator' }] }, 0]
}, {
    'x': [1,2,3,4,5]
}) // 15

In this example, we run the reduce operation on a variable called "x", and we set up instructions to add the "current" value to the "accumulator", which we have set to 0.

Similarly, you can also do map operations:

logic.run({
    'map': [[1,2,3,4,5], { '+': [{ var: '' }, 1] }]
}) // [2,3,4,5,6]

If var is left as an empty string, it will assume you're referring to the whole variable that is accessible at the current layer it is looking at.

Example of a map accessing variables of the objects in the array:

logic.run({
    'map': [{var : 'x'}, { '+': [{ var: 'a' }, 1] }]
},
{
    'x': [{ a: 1 }, { a: 2 }, { a: 3 }, { a: 4 }]
}) // [2,3,4,5]

You can easily nest different operations in each other, like so:

logic.run({
    max: [200, {
        '*': [12, {var: 'a' }]
    }]
}, {
    a: 16
}) // 200

The engine also supports Handlebars-esque style traversal of data when you use the iterative control structures.

For example:

logic.run({
    'map': [{var : 'x'}, { '+': [{ var: 'a' }, { var: '../../adder'}] }]
},
{
    'x': [{ a: 1 }, { a: 2 }, { a: 3 }, { a: 4 }],
    'adder':  7
}) // [8, 9, 10, 11]

---

Migrating from json-logic-js

This library is designed as a drop-in replacement for json-logic-js. Most existing rules should work without modification, while giving you access to all the performance and feature improvements mentioned above.