objex

Cloud storage explorer that runs entirely in the browser. Connect to S3, Azure, GCS, R2, MinIO -- browse files, query data with SQL, and visualize geospatial formats on interactive maps. No backend required.

graph LR
    Cloud["Cloud Storage<br/>S3 / Azure / GCS / R2"] -->|HTTP range requests| App["objex<br/>(browser SPA)"]
    App --> Browse[File Tree]
    App --> Query["DuckDB-WASM<br/>SQL Engine"]
    App --> Map["MapLibre + deck.gl<br/>Geo Visualization"]
    App --> View["18+ Viewers<br/>Tables, Code, Notebooks, PDF, 3D..."]

Loading

Features

• Browse cloud storage (S3, GCS, Azure, R2, B2, DigitalOcean, Wasabi, Storj, Hetzner, Contabo, Linode, OVHcloud, MinIO, direct URLs)
• Query Parquet, CSV, JSONL with SQL (DuckDB-WASM, cancellable queries)
• Visualize GeoParquet, GeoJSON, COG, PMTiles, FlatGeobuf, Zarr on maps (MapLibre + deck.gl)
• View 100+ file formats: code (30+ languages), Jupyter notebooks, PDF, 3D models, archives, media
• Share via URL -- ?url=<storage-url>#<view> encodes full viewer state
• i18n -- English + Arabic with automatic RTL layout
• Zero backend -- everything runs client-side

Supported Formats

Category	Formats
Tabular	Parquet, CSV, TSV, JSONL, NDJSON
Geo vector	GeoParquet, GeoJSON, Shapefile, GeoPackage, FlatGeobuf
Geo raster	COG, PMTiles, Zarr v2/v3
Point cloud	COPC, LAZ, LAS
Notebooks	Jupyter (.ipynb), marimo
Code	30+ languages (Python, TS, Rust, Go, SQL...)
Documents	Markdown, PDF, text, logs
Media	Images, video, audio
3D	GLB, glTF, OBJ, STL, FBX
Archives	ZIP, TAR, GZ, 7Z, RAR
Database	DuckDB, SQLite

npm Packages

Two packages are published for downstream use:

@walkthru-earth/objex -- Full Svelte 5 Library

Components, stores, and utilities for building geospatial storage explorers.

npm install @walkthru-earth/objex

import { parseStorageUrl, formatFileSize } from '@walkthru-earth/objex';
import { UrlAdapter } from '@walkthru-earth/objex/storage';
import { parseWKB } from '@walkthru-earth/objex/utils/wkb';
import { buildGeoArrowTables } from '@walkthru-earth/objex/utils/geoarrow';
import { readParquetMetadata } from '@walkthru-earth/objex/utils/parquet-metadata';
import { getFileTypeInfo } from '@walkthru-earth/objex/file-icons';

Requires svelte ^5 and @sveltejs/kit ^2 as peer dependencies. Heavy deps (DuckDB, deck.gl, MapLibre, Arrow, hyparquet) are optional peers -- only install what you need.

@walkthru-earth/objex-utils -- Pure TypeScript Utilities

Zero Svelte dependency. Works with any JS framework or Node.js.

npm install @walkthru-earth/objex-utils

import {
  parseStorageUrl,
  parseWKB,
  buildGeoArrowTables,
  readParquetMetadata,
  getFileTypeInfo,
  formatFileSize,
  generateHexDump,
  classifyType
} from '@walkthru-earth/objex-utils';

Exports

Export path	What
@walkthru-earth/objex	All types, utils, storage, query engine
./storage	StorageAdapter, UrlAdapter
./query	QueryEngine, QueryCancelledError
./utils/wkb	parseWKB, toBinary, findGeoColumn
./utils/geoarrow	buildGeoArrowTables, normalizeGeomType
./utils/storage-url	parseStorageUrl, looksLikeUrl
./utils/parquet-metadata	readParquetMetadata, extractEpsgFromGeoMeta
./utils/format	formatFileSize, formatDate, formatValue, getFileExtension, jsonReplacerBigInt
./utils/hex	generateHexDump
./utils/column-types	classifyType, typeColor, typeBadgeClass
./file-icons	getFileTypeInfo, getDuckDbReadFn, getViewerKind
./types	FileEntry, Connection, Tab, WriteResult, Theme

The main export also includes copyToClipboard, handleLoadError, and shared constants (WGS84_CODES, STORAGE_KEYS, DEFAULT_TARGET_CRS, etc.).

Quick Start (Development)

pnpm install
pnpm dev          # http://localhost:5173

Contributing

See CONTRIBUTING.md for development setup, architecture, geospatial pipeline docs, and viewer checklist.

License

CC BY 4.0 -- hi@walkthru.earth