WherobotsDB The spatial compute engine built for physical-world data WherobotsDB is the distributed spatial engine with full geometric correctness at any scale. 300+ spatial functions for vector and raster data, with native Spark SQL for tabular operations. Built by the original creators of Apache Sedona. START BUILDING FREE 300+ spatial functions Vector, raster, topology, and spatial indexing. 3x faster, 46% lower cost Spatial joins that run in minutes, not days. From city block to the entire globe Horizontally distributed on Apache Spark. Scales with your data. The benchmark leader in spatial compute WherobotsDB delivers the fastest and most capable engine for cost-efficient performance across general-purpose spatial SQL workloads. The following table was run across multiple compute engines using the Apache Sedona SpatialBench at Scale Factor 1000. In a separate test, WherobotsDB completed a Texas-statewide raster spatial join in 3 minutes 28 seconds. BigQuery timed out. Scale Factor 1000 Query Capability Matrix SpatialBench Query # WherobotsDB Sedona on Serverless Spark Proprietary SQL Serverless Proprietary Spark Clusters with Spatial SQL Sedona on Managed Spark Clusters 1: Spatial filter, aggregation, sorting 2: Spatial filter, aggregation, sorting 3: Spatial filter, aggregation, sorting 4: Lightweight spatial join and aggregation 5: Spatial aggregation 6: Heavy spatial join 7: Geometry construction and access 8: Heavy distance join 9: Heavy polygon self spatial join 10: Heavy spatial left join 11: Multi-way spatial join 12: KNN join SpatialBench for Apache Sedona Key capabilities Unlimited Scale Scale workloads seamlessly from small to a planetary scale. Apache Spark & Sedona Compatible Migrate existing Apache Spark and Sedona workloads without changing a single line of code. Fully compatible out-of-the-box. Nothing to Manage But Your Code Serverless by default, or deploy securely in your own VPC. 300+ Spatial Functions for Vector and Raster Data High performance spatial joins with 300+ ST_ and RS_ functions built to accelerate productivity. Low Cost + High Performance Spatial Operations WherobotsDB delivers fast, cost-efficient performance across spatial SQL and python workloads*. WherobotsDB benchmarks at 3x faster with up to 45% better price-performance than comparable engines, and at lower cost than PostGIS at scale. *This represents actual cost benchmarking results for SpatialBench queries 7-12 with a scale factor of 1000 across three of the most comparable deployment options. Contact us to compare all benchmarks WherobotsDB Lowest Proprietary SQL Serverless Engine x2.92 higher Apache Sedona on Serverless Spark Engine x2.03 higher Proprietary Cloud Managed Spark x2.52 higher Relative cost across operations that completed Workload Flexibility From Small to Planetary Scale Workloads Perform city to global scale spatial data operations in a single solution. City Region Country Global Productivity Accelerate Innovation Economically SQL and Python 300+ ST_ and RS_ Functions Raster Vector Tabular SPATIAL JOIN Data Products “Spark is incredibly powerful — but it’s also a huge learning curve, Wherobots shortened the painful part of Spark and gave us production-grade scalability without having to babysit clusters.” Ben Hudson Co-founder and Head of Applied Science, aarden.ai Apache Spark & Sedona Compatible Lift and shift your existing Spark + Sedona workloads with the same APIs and code, now with improved performance and lower cost. Learn More scalaspark.read.format("csv") .option("header", "true") .load("cities.csv") .createOrReplaceTempView("cities"); val result = spark.sql( "SELECT * FROM cities WHERE ST_Contains(area, point)" ) result.show() Deployment & Compatibility Zero Migration. Total Flexibility. Keep your existing Apache Spark or Sedona code, run it serverless or in your VPC Serverless by Default Zero infrastructure to manage Auto-scaling, cost-efficient Works out-of-the-box Learn more Deploy in Your Cloud (BYOC) Full control in your VPC Enterprise-ready security Works behind your firewall Reach out to our team Go deeper WherobotsDB is 3x Faster with up to 45% Better Price Performance The next generation of WherobotsDB delivers measurable gains across spatial SQL workloads, with benchmarks to prove it. LEARN MORE Raster Processing at Scale: Out-of-Database Architecture How WherobotsDB processes terabyte-scale satellite imagery, elevation models, and sensor data without storing raster data inside the database. LEARN MORE Streaming Spatial Data with Spark Structured Streaming How to build real-time spatial pipelines that ingest and process GPS and IoT data as it arrives, without the workarounds traditional systems require. LEARN MORE Get Started with WherobotsDB Explore the platform, build powerful solutions, or automate your geospatial workflows. Discover Seamless access to the right datasets, AI models and solution accelerators for your use case. LEARN MORE Build Start using WherobotsDB with familiar Apache Spark/Sedona syntax. LEARN MORE Automate Orchestrate spatial data flows at scale with Apache AirFlow or our REST API. LEARN MORE FAQ How does WherobotsDB compare to other geospatial analytics engines? WherobotsDB is a distributed geospatial analytics engine with full geometric correctness at planetary scale. On the Apache Sedona SpatialBench at Scale Factor 1000, WherobotsDB completed all 12 spatial queries, including heavy spatial joins that the next best Spark-based engine with Spatial SQL could not finish within a 10-hour timeout. WherobotsDB benchmarks at 46% lower cost than that engine and delivers up to 3x faster performance on analytical queries. For raster-vector workloads, WherobotsDB completed a Texas-statewide raster spatial join in about 3.5 minutes. BigQuery timed out on the same test. Which spatial functions are supported in Wherobots? WherobotsDB includes 300+ spatial functions for vector and raster data, covering geometry processing, spatial predicates, coordinate reference systems, and large-scale spatial joins. Advanced functions include kNN joins for nearest neighbor queries, zonal statistics on raster layers, travel isochrones for accessibility analysis, GPS map matching for mobility data, and PMTiles generation for map visualization. Native Spark SQL handles tabular operations alongside the spatial functions. What raster data processing functions does WherobotsDB support? WherobotsDB includes 100+ raster RS_ functions for processing satellite imagery, elevation models, and sensor data at scale. These cover in-database and out-of-database raster loading, affine transformations (translation, scaling, rotation, shearing), union aggregation to combine multiple rasters into multi-band outputs, zonal statistics, map algebra, raster tiling, and output writers for GeoTIFF, ArcGrid, and PNG formats. WherobotsDB’s out-of-database raster architecture processes terabyte-scale imagery without moving data into the database, enabling planetary-scale inference and zonal statistics. Why is WherobotsDB faster than general-purpose engines for spatial data analysis? WherobotsDB is built from the ground up for spatial data workloads. The engine uses a Rust-native, Arrow-columnar execution layer with zero-copy geometry handling via GeoArrow, optimized at the compute, query planner, and function levels. Production teams report processing times dropping from 39 days to under one day (GeoPostcodes) and from 7 days to 30 minutes (Aarden.ai). WherobotsDB separates compute from storage, scales horizontally on Apache Spark, and maintains 100% code compatibility with open-source Apache Sedona. Get started WherobotsDB is part of Wherobots, the AI Context Engine for the Physical World. TALK TO US START BUILDING FREE
WherobotsDB is 3x Faster with up to 45% Better Price Performance The next generation of WherobotsDB delivers measurable gains across spatial SQL workloads, with benchmarks to prove it. LEARN MORE
Raster Processing at Scale: Out-of-Database Architecture How WherobotsDB processes terabyte-scale satellite imagery, elevation models, and sensor data without storing raster data inside the database. LEARN MORE
Streaming Spatial Data with Spark Structured Streaming How to build real-time spatial pipelines that ingest and process GPS and IoT data as it arrives, without the workarounds traditional systems require. LEARN MORE
