@t0mer/n8n-nodes-latlng

An n8n community node for LatLng. It covers forward and reverse
geocoding, place search, nearby places, autosuggest and place categories. It also renders static map
images and fetches vector tiles and TileJSON, all based on OpenStreetMap data.

Unofficial. This package is not affiliated with, endorsed by or supported by LatLng.

• Demo
• Installation
• Credentials
• Operations
• Coordinates
• Options
• Quota and errors
• Attribution and usage terms
• Development

Demo

▶ Watch the demo video (MP4, 43 s)

The workflow is in demo/workflow.json:

1. Geocoding → Forward looks up "Dizengoff Square, Tel Aviv".
2. Place → Nearby lists 5 cafés within 500 m of it.
3. Static Map → Get Image draws the square (red) and the cafés (blue).

Import the workflow and pick your LatLng credential in each LatLng node.

Installation

From the n8n UI (self-hosted n8n):

1. Go to Settings → Community Nodes → Install.
2. Enter @t0mer/n8n-nodes-latlng and confirm.

Manually (e.g. Docker or queue mode):

cd ~/.n8n/nodes
npm install @t0mer/n8n-nodes-latlng
# restart n8n

See the n8n docs on installing community nodes.

Credentials

Create a LatLng API credential. You get keys from the LatLng dashboard at
dash.latlng.work.

• Put the server key in the API Key field. A Maps key (pk_latlng_…) there makes the server-only
  endpoints answer 403.
• Tile and Dataset operations fail with a clear error if the Maps key is empty. Dataset tiles accept
  only a Maps key that belongs to the dataset owner.
• Test on the credential calls GET /v1/places/categories once, which counts against your quota.

Operations

Every operation's description says what it returns and in which units. The node is also marked as
usable as an AI agent tool.

Geocoding

Example output (Forward, Simplify on, one item per match). Without Language, names come back
in the local language:

{
  "name": "כיכר דיזנגוף",
  "lat": 32.0779938,
  "lon": 34.7743493,
  "osm_type": "R",
  "osm_id": 13645250,
  "osm_value": "square",
  "type": "locality",
  "city": "תל אביב–יפו",
  "country": "ישראל",
  "countrycode": "IL",
  "extent": [34.7738809, 32.0782606, 34.7745159, 32.077723]
}

Reverse returns the same fields for the address at the point.

Place

Example output (Nearby, one item per place, distance_m in meters):

{ "id": "13516609157", "name": "Plaza Café", "category": "cafe", "lat": 32.0775717, "lon": 34.7739451, "distance_m": 36.9 }

Example output (Autosuggest, one item per suggestion):

{ "name": "דיזנגוף", "type": "address", "category": "street", "lat": 32.0765505, "lon": 34.7746435, "distance_m": 162, "country": "IL" }

Get Categories outputs one item per category, e.g. { "category": "cafe", "osm_tag": "amenity:cafe" }.
Use the category value in the Category filter.

Static Map

• Framing: Center and Zoom (latitude, longitude, zoom 0–20), or Bounding Box (min/max latitude
  and longitude; the zoom is fitted automatically).
• Style: light, dark (default), grayscale, black, white or contrast.
• Markers: latitude, longitude, color and a short label for each pin.
• Paths: a line through two or more points, with weight (pixels), color and opacity (0–1).
• GeoJSON Overlay: any GeoJSON object to draw on top of the map.

Current API behavior (checked September 2026): the live API draws markers in the chosen color,
but it does not draw marker labels, paths or GeoJSON overlays. The request still succeeds; those
parts are just missing from the image. The node sends them as LatLng documents them, so they
should start working if the API adds support.

The JSON part of the output holds the request parameters (never the key) plus fileName, mimeType
and fileSize. If the request gets very long, for example with hundreds of markers, the API may reject
it. The error then suggests fewer markers or a simpler GeoJSON.

Tile and Dataset (Maps key required)

• TileJSON echoes your key inside the tiles URLs, so the node replaces it with key=REDACTED in the
  output.
• Tile coordinates are checked before the call: z, x and y must be whole numbers, with
  0 ≤ x, y < 2^z.

Coordinates

You always enter Latitude and Longitude as two separate number fields in decimal degrees. The
node checks the ranges (latitude −90 to 90, longitude −180 to 180) before calling the API.

You never need to know the order a particular endpoint expects, because the node handles it:

• Query parameters use lat/lon.
• GeoJSON uses [lon, lat]. With Simplify on, results come back as flat lat/lon fields.
• Static maps use lng,lat.

Options

Each output item is paired with its input item. With Continue On Fail, a failed item becomes
{ "error": "…", "statusCode": 401 } and the rest of the batch keeps running.

Quota and errors

• LatLng uses one shared quota for all calls: the free plan allows 3,000 calls per day. Every
  call counts, including tiles, static maps and the credential test.
• 429 and 5xx responses are retried at most twice (after 1 s and 3 s). Other errors are not
  retried, because every retry costs quota.

API keys never appear in node output or error messages.

Attribution and usage terms

Map data © OpenStreetMap contributors, available under the
ODbL. If you publish a static map image, show this
attribution visibly near it.

Your use of the API is subject to LatLng's terms and plan limits. See latlng.work.

Development

npm install
npm run lint     # n8n-node lint (strict)
npm run build
npm test         # unit tests, no network
npm run dev      # n8n with this node loaded, at http://localhost:5678

npm run smoke calls each operation once against the real API, using keys from a local .env file
(LATLNG_API_KEY, and optionally LATLNG_MAPS_KEY). It writes sanitized responses to
test/fixtures/live/. It makes about 10 calls, so don't run it in a loop. Set LATLNG_DATASET_ID to also check a dataset.

Releases are published by pushing a YYYY.M.PATCH tag. See .github/workflows/publish.yml.

License

MIT