Your AI can now contextualize physical world data using Wherobots Spatial AI Coding Tools Learn More
Menu
TRY WHEROBOTS
TRY WHEROBOTS
Your AI can now contextualize physical world data using Wherobots Spatial AI Coding Tools Learn More

Discover

The Global Hub for physical-world data and AI

One place for every physical-world AI asset: geospatial datasets, AI models, notebooks, workflows, and solution recipes. The organizational memory that makes your team and your AI agents more productive with every build.

START BUILDING
Curated datasets, ready to query
Overture Maps Foundation, Foursquare, and hosted spatial datasets. Bring your own from S3, GCS, Azure, or Databricks Unity Catalog.
Pre-trained AI models for Earth intelligence
Object detection, segmentation, classification, and map matching. Bring your own PyTorch model or start with the built-in library.
Recipes, not blank pages
Solution notebooks for real-world use cases. Start from working code, deploy to production.

Geospatial datasets, AI models, and solution notebooks in one hub

Solution Notebooks

Working code for insurance risk, agricultural field detection, mobility analytics, and infrastructure mapping. Start from a proven example.

Overture Maps and Foursquare Datasets

Places, buildings, transportation networks, and administrative boundaries from Overture Maps Foundation and Foursquare. Ready to query in WherobotsDB.

Bring Your Own Data

Connect S3, GCS, Azure, or Databricks Unity Catalog. Register STAC collections. Upload GeoParquet or raster tiles. Your data stays in your storage.

RasterFlow Models

Built-in models for field boundary detection, canopy height estimation, road detection, and urban infrastructure mapping. Or bring your own PyTorch model.

Share Across Your Team

Share notebooks, geospatial datasets, and model configurations across your organization. Build on each other’s work.

RasterFlow Models for Satellite Imagery Analysis

Pre-trained computer vision models for common Earth observation tasks. Run inference at any scale with RasterFlow, or bring your own PyTorch model.

Object detection

Detect buildings, solar panels, vehicles, and infrastructure in satellite or aerial imagery.

LEARN MORE

Segmentation

Delineate boundaries for crop fields, water bodies, and vegetation areas from satellite imagery.

LEARN MORE

Classification

Categorize land cover types, urban areas, and vegetation classes across regions or entire continents.

LEARN MORE

Map Matching

Snap noisy GPS data to road segments to simplify analysis of mobility data.

LEARN MORE
Integrations & Catalogs

Connect your geospatial datasets from anywhere

Hosted geospatial datasets from Overture Maps and Foursquare, your own cloud storage, Databricks Unity Catalog, or any STAC catalog. The Global Hub connects to all of them.

Hosted Datasets

Popular, curated datasets built-in to Wherobots for seamless use.

High quality global datasets from Overture and Foursquare with places, buildings, transportation infrastructure and more.

Learn More

Your Cloud Storage

Connect AWS S3, Google Cloud Storage, or Azure Blob Storage. Query geospatial datasets directly from WherobotsDB without migration.

Learn More

Databricks Unity Catalog

Integrate with Databricks Unity Catalog.

Register Unity as a catalog source. Browse schemas, tables, and views.

Learn More

STAC Collections

Build references to Spatial Temporal Asset Catalog (STAC) collections

Register STAC URLs to index imagery and metadata for search & inference.

Learn More

Learning Resources

VIEW MORE
Next Steps

Explore the
Wherobots Platform

Build

Build spatial AI with SQL, Python, and the Spatial AI Coding Assistant.

LEARN MORE

Automate

Take spatial data pipelines to production with Apache Airflow and Job Runs.

LEARN MORE

AI for Earth

Run satellite imagery analysis at any scale with RasterFlow.

LEARN MORE

FAQ

Which geospatial datasets are built into Wherobots?

Global Hub includes geospatial datasets from Overture Maps Foundation and Foursquare, covering places, buildings, transportation networks, and administrative boundaries globally. These datasets are pre-loaded and ready to query in WherobotsDB. Teams can also connect their own geospatial datasets from S3, GCS, Azure, or Databricks Unity Catalog without migration.
For more information, see the documentation.

Which AI models are available in Wherobots?

The Global Hub provides pre-trained computer vision models with RasterFlow. Built-in models include Fields of the World for agricultural field boundary detection, Meta CHM v1 for canopy height estimation, Tile2Net for urban infrastructure detection, and ChesapeakeRSC for rural road detection. RasterFlow supports semantic segmentation, regression, and patch-based processing. You can also bring your own PyTorch model using the Machine Learning Model (MLM) standard.

How do I connect my own cloud storage?

Wherobots connects to AWS S3 buckets using cross-account IAM roles, maintaining your existing governance and permissions. Google Cloud Storage and Azure Blob Storage are also supported. Connect your buckets and query geospatial datasets directly from WherobotsDB without migration. The Global Hub serves as the central geospatial data management layer for all connected data sources, including your own cloud storage, Databricks Unity Catalog, and STAC catalogs.

How do I access aerial imagery available through a SpatioTemporal Asset Catalog (STAC)?

WherobotsDB includes a built-in STAC reader that connects to any SpatioTemporal Asset Catalog. Load STAC items and collections directly into Sedona DataFrames using Python or Scala. Connect to STAC catalogs like Element84 Earth Search for Sentinel-2 imagery, or register your own STAC endpoints including private catalogs with API key authentication. Once loaded, process the data with WherobotsDB’s 300+ spatial functions or write results to GeoParquet for storage in your lakehouse.
Sphere

Get started

Wherobots is the AI Context Engine for the Physical World.

TALK TO US START BUILDING