Plus Codes Converter

The structure of a Plus Code

A full Plus Code is a string of 10–15 characters with a + separator inserted after position 8. Each character is one of 20 base-20 digits drawn from the alphabet 23456789CFGHJMPQRVWX — chosen specifically to avoid look-alikes (no I/L/1, O/0).

87G8 P2X7 + 9P
│  │ │  │   │
│  │ │  │   └ pair 5 — 4×5 grid refinement (1 m–14 m cell)
│  │ │  └ pair 4 — 1/400° × 1/400° (~275 m)
│  │ └ pair 3 — 1/20° × 1/20° (~5.5 km)
│  └ pair 2 — 1° × 1° (~111 km)
└ pair 1 — 20° × 20° worldwide grid

The first four pairs locate the point through successive twenty-fold subdivisions of latitude and longitude. The fifth pair (and any beyond) refines through a 4-wide × 5-tall grid: each character splits the previous cell into 20 sub-cells arranged in 4 columns and 5 rows. This is what gives the cell its slight east–west / north–south aspect ratio at high precision.

Precision is a length, not a separate field

Like MGRS, Plus Codes encode precision into the string itself. Trim characters from the right to coarsen; add to refine. The default 10-character code lands at the 14 m cell, which is enough for "a specific building" in most contexts.

How Plus Code math works

The encoding is pure base-20 in latitude and longitude, with no projection involved. Plus Codes ride on plain WGS-84 lat/lon (the same coordinate frame the GPS satellite system uses) and assume the equirectangular plane — they don't correct for Earth's curvature inside each cell. That's a feature for human use: a Plus Code is a fixed area in degrees, not in metres.

Ten worked examples — both directions

The Gulf-of-Guinea row is the cleanest illustration of the base-20 indexing. At exactly 0°N, 0°E the cell sits at the SW corner of every refinement level — every pair after pair 1 lands on the first letter of the alphabet (2), so the whole tail of the code is 22222+22. The North-Pole row is the cleanest illustration of the system's global coverage: unlike UTM and MGRS, Plus Codes have no zone exception above 84°N or below 80°S. The cell's east-west width does shrink toward zero at the poles, but the code itself remains valid.

Misconceptions worth getting straight

"A Plus Code is a point"

No — it's a cell. A 10-character Plus Code names a 14 m × 14 m area. The lat/lon you decode is the cell's centre, but any point inside the cell shares the same code at that length. For precise point work, decode to lat/lon and use the lat/lon downstream.

"Plus Codes and what3words are the same thing"

Different. Plus Codes are an open standard (Apache 2.0) with a deterministic, offline-computable encoding. what3words is a proprietary system using a 40,000-word dictionary mapped to a 3 m × 3 m grid by the company's closed algorithm; using it commercially requires a paid licence. Plus Codes degrade gracefully with precision; what3words is fixed at 3 m. Plus Codes work offline; what3words requires an API call.

"The first letter tells you the continent"

Roughly, but not reliably. The first character splits the world into a 20° latitude band, the second into a 20° longitude column. There are 18 latitude bands (because lat ranges from -90° to +90°) and 18 longitude columns (from -180° to +180°). Each two-character pair therefore identifies one of about 162 macro-cells worldwide — about the size of Texas. A given continent occupies many first-pair values, and a given first-pair value may straddle multiple continents.

"Plus Codes work anywhere Google Maps works"

Plus Codes work everywhere, including places Google Maps doesn't cover well: dense informal settlements, humanitarian operations zones, polar research stations. That was the original design intent. Google Maps recognises them as search inputs, but the codes themselves exist whether Google's app is involved or not — any Apache-2.0-licensed implementation produces identical codes.

"Longer codes are always better"

Not for human use. A 10-character code reads cleanly aloud and survives transcription errors. A 15-character code is surveying-grade but no human will read it back without stumbling. Match precision to use case: 10 chars for everyday-address replacement, 11–12 for delivery to a doorway, 13–15 only for engineering or surveying.

When to use Plus Codes, when not to

How to verify a Plus Code

Three sanity checks any Plus Code result should pass:

1. The + separator is at position 9. Always exactly 8 characters before the plus, then the plus, then 1–7 characters after. Any other placement is malformed.
2. Every character is in the 20-letter alphabet.If you see an I, L, O, 1, or 0, the code has been corrupted — usually by autocorrect or by a font that confuses characters.
3. Round-trip the conversion. Convert Plus Code → lat/lon → Plus Code at the same length and confirm it matches. The encoding is deterministic, so any disagreement is a sign of a malformed input.

How this converter is built

Encoding and decoding run entirely in the browser via src/lib/coords/plus-codes.ts, a from-scratch implementation of Open Location Code v1.0 (no Google library dependency). The input is auto-detected as Plus Code or lat/lon by checking for a + at position 9 with every character in the OLC alphabet; anything else is parsed as lat/lon. The deep report below the map shows the cell bounds in degrees and metres, the same point in all six common notations, the nearest place (Mapbox v6 server-side proxy with no-store cache), and the elevation (USGS 3DEP for US, OpenTopoData SRTM30m elsewhere) — per CLAUDE.md §19.2 and §19.4. No coordinates are ever logged or retained.