How much accuracy do I lose by rounding coordinates?

At the equator, 5 decimal places is about 1.1 metres, 6dp is about 0.11 metres, and 4dp is about 11 metres. Web maps rarely need more than 5dp, so trimming from the typical 15 stored digits is lossless for display purposes.

Does minifying change the geometry?

Stripping whitespace never changes anything. Rounding shifts each coordinate by at most half a unit of the chosen precision, which is invisible on a normal map but can matter for survey-grade data, so keep an unrounded copy if exact values are required.

Does it validate the GeoJSON?

It parses the input as JSON and rejects invalid syntax with an error message. It does not enforce the full GeoJSON spec, but it does walk every coordinate array so malformed geometry is usually surfaced.

Is my data uploaded anywhere?

No. Parsing, rounding, and re-serialising all happen locally in your browser, so sensitive location data never leaves your machine.

Why is my file still large after minifying?

Most of a GeoJSON file's size is the coordinate digits themselves, not whitespace. If you only strip whitespace the saving is modest; rounding to 5 or 6 decimals usually removes the bulk of the bytes.

GeoJSON Minifier — Gera Tools

A GeoJSON minifier makes map data files smaller so they download and render faster in the browser. GeoJSON exported from GIS tools is often pretty-printed with indentation and stores coordinates with 15 decimal places of floating-point precision — far more than any web map can display. This free tool removes that overhead client-side, with no upload and no account.

How it works

Two independent transforms run on your data:

Whitespace removal. The input is parsed with JSON.parse and re-serialised with JSON.stringify and no spacing argument, which collapses all indentation and newlines.
Coordinate rounding. The tool recursively walks every array in the document. A coordinate pair is detected as an array whose entries are all numbers; each number is rounded to your chosen decimal precision using Math.round(n * 10^d) / 10^d. Trailing zeros are dropped automatically by JavaScript’s number-to-string conversion, which is where most of the byte savings come from.

Rounding precision maps to real-world distance at the equator like this: 4dp is roughly 11 m, 5dp roughly 1.1 m, and 6dp roughly 0.11 m. Longitude precision shrinks toward the poles, so the figures are an upper bound.

Tips and notes

For street-level web maps, 5 decimal places is the sweet spot — typically a 50–70% size reduction with no visible change.
Keep an unrounded master copy if you need survey-grade accuracy; rounding is lossy and not reversible.
Combining this with gzip on your server compounds the saving, because shorter, more repetitive numbers compress better.
The tool handles nested geometries (MultiPolygon, GeometryCollection) because it walks the whole tree rather than assuming a flat structure.