What is an encoded polyline?

An encoded polyline is a compact string that stores a series of coordinates, used by Google Maps and many routing APIs to transmit a path efficiently instead of a long array of numbers.

Why are deltas encoded instead of absolute values?

The algorithm stores the difference between each point and the previous one. Because consecutive points are close, the deltas are small and compress to fewer characters than full coordinates.

What does precision 5 versus 6 mean?

Precision 5 multiplies coordinates by 100000, the Google default, giving about a metre of resolution. Precision 6 multiplies by 1000000 for finer detail and is common in OSRM output. The decoder must use the same precision.

How are negative coordinates handled?

Each delta is shifted left by one bit, and negative values are bitwise inverted. This signed encoding lets the rest of the algorithm treat every value as a chunked base-64 number.

Can I round-trip the result?

Yes. Feed the output into the Google Polyline Decoder at the same precision and you recover the original points, subject only to the rounding implied by the chosen precision.

Google Polyline Encoder

Encoded polylines

A route or shape is naturally a list of coordinates, but sending hundreds of floating-point numbers is wasteful. The Google encoded polyline format squeezes that path into a single short string that map and routing APIs understand directly.

How it works

Each coordinate is multiplied by ten to the precision (100000 for precision 5) and rounded to an integer. The algorithm then stores the delta from the previous point. Each delta is left-shifted by one bit, inverted if negative, split into 5 bit chunks, and each chunk has 0x20 OR-ed in while more chunks follow before adding 63 to land in printable ASCII.

[38.5, -120.2], [40.7, -120.95], [43.252, -126.453]
  ->  _p~iF~ps|U_ulLnnqC_mqNvxq`@

The first point is encoded as a delta from zero, so the string is fully self-contained.

Example and tips

Always match the precision when decoding, or the coordinates come out scaled wrongly. Order matters: the points must be listed in path order because the encoding is delta-based. If your data comes from OSRM, Valhalla or another high-precision engine, select precision 6; for the Google Directions API use precision 5.