Index your own datasets

Bring GeoJSON, CSV, and Parquet into E2 and keep a portable copy for downstream tools.

Documentation navigation

Start with E2

SDK guides

CLI guides

MCP guides

Applications

Framework quickstarts

Upload and index

terminal

e2 storage upload ./restaurants.geojson
  --db ./e2.db --connect-to restaurant
  --out ./restaurants.geoparquet

Uploads require a feature connection through --feature, --connect-to, --feature-id, or row metadata. Rows without a valid feature connection are rejected.

Input formats

Format	Expected coordinates
`.geojson / .json`	GeoJSON geometry or lat/lng fields.
`.csv`	lat + lng or latitude + longitude columns.
`.parquet / .geoparquet`	Normalized E2 fields or coordinates and feature metadata.

The storage layer assigns feature_id, feature, e2_id, geometry, source metadata, and observation time. It preserves input properties.

Import and inspect

terminal

# Import a published Parquet file into a local database
e2 import-file ./published.parquet --db ./e2.db --region e2r:region:nyc

# Inspect rows and feature counts
e2 query --db ./e2.db --window all --limit 100
e2 stats --db ./e2.db

All developer interfaces