About M33
M33 builds Fabric — the tools that take raw Earth observation data and make it analysis-ready, so the people who work with it spend their time on the analysis, not the preparation.
Why We Exist
The bottleneck in Earth observation moved years ago — from getting the data to preparing it. Before a GIS analyst can answer a single question about vegetation, flooding, or land-use change, they download tiles, reconcile coordinate systems, align pixel grids, normalize missing values, and resolve the differences between sensors that were never meant to be compared. Analysts routinely spend 45 to 80 percent of a project not analyzing data, but preparing it.[1]CrowdFlower 2016 / Anaconda 2020forbes.com ↗
And the knowledge they build doing it — which bands to use, what order to run things in, how to handle the edge cases — stays trapped in one-off scripts and notebooks. Every new project starts from scratch. When an analyst leaves, the workflow leaves with them.
That preparation layer is what M33 builds.
What We Build
The runtime that turns multi-source Earth observation data into harmonized, analysis-ready outputs — reproducibly, the same way every time. Every result carries a complete provenance record of what ran, on what data, in what order.
LiveBuild processing workflows without writing code, and watch them run. Compose a workflow once and it becomes something you can reuse, hand off to a teammate, and audit later.
LiveDescribe the analysis you need in plain language, and Iris turns it into a ready-to-run workflow — so you can go from question to pipeline without building it by hand.
In developmentPrinciples
Outputs go to storage you control. Results are delivered as standard formats (Cloud-Optimized GeoTIFF, GeoJSON, STAC catalogs) that work with every major GIS tool. We do not build walls around your data or your workflows. Data does not accumulate on M33 infrastructure.
Every output includes a provenance record: which inputs were used, what parameters were applied, what processing steps ran, in what order. In a field where wildfire response, agricultural policy, and carbon-credit verification depend on satellite-derived evidence, "trust us" is not good enough.
We are not replacing ArcGIS, QGIS, Google Earth Engine, or any data provider. We fill the gap between raw data and analysis-ready data. If your existing tools work for the analysis step, Fabric makes sure the data is ready for them.
The Team
M33 was founded by Michael Keys.
Our advisory council includes GIS domain experts who test our workflows against real-world projects and keep us honest about what matters in the field.
We are building the team. If you work in geospatial processing, remote sensing, or Earth observation infrastructure and what you have read here resonates, .
Let's talk