Guides
Running Earth Engine scripts
Check a Google Earth Engine Code Editor script against Atlas, run it, and share the resulting report.
Atlas runs Earth Engine Code Editor scripts as they are. A run produces a report: the script, everything it printed, each chart’s data table and every map layer, stored as a graph so tiles render on demand.
Check before you run #
POST /ee/check reads the script without running it and lists the ee.* calls and datasets it uses, with anything unsupported named:
Terminal
curl -X POST "$ATLAS_API/ee/check" \
-H "Content-Type: application/json" \
-d '{"script": "var img = ee.Image(\"USGS/SRTMGL1_003\"); print(img.bandNames());"}'
GET /ee/datasets lists the collection ids Atlas knows and whether each is available now or needs-ingest.
Run it #
Running needs a signed-in user:
Terminal
curl -X POST "$ATLAS_API/ee/run" \
-H "Authorization: Bearer $ATLAS_TOKEN" \
-H "Content-Type: application/json" \
-d '{"title": "Seoul elevation", "script": "…", "public": false}'
The response has the report id. Open GET /ee/reports/{id}/view for the map, charts and output, fetch layer tiles at /ee/reports/{id}/layers/{i}/{z}/{x}/{y}.png, or export the report to kepler.gl with /ee/reports/{id}/kepler.json.
Needs a long-running server
Script execution needs a long-running server. A deployment without it answers
POST /ee/run with 501; checking, datasets and saved reports still work.
Your own tables #
Upload a GeoJSON FeatureCollection with PUT /ee/assets under projects/<you>/assets/<name>, then load it in a script with ee.FeatureCollection('projects/<you>/assets/<name>').