Point your phone. Get lab-grade particle size distribution in seconds. No lab. No waiting. No expensive equipment. grai replaces sieve analysis on-site.
The EU-funded research project GRID — led by our partner university BOKU (University of Natural Resources and Life Sciences, Vienna) — has launched a Kaggle competition challenging participants to predict soil grain size distributions from photographs. The competition is built on the original dataset used to develop the first version of grai, the same AI technology that powers deepsoil today. This is an exciting opportunity for data scientists and machine learning enthusiasts to push the boundaries of photogranulometry.
Join the Competition on Kaggle →Replace your next sieve analysis with a simple smartphone workflow.
Fill the photobox with your soil sample, position your phone, and take a photo. That's it — no special equipment beyond your smartphone and our photobox.
Upload the image to grai. Our AI processes it in seconds, segmenting and measuring every visible particle using deep learning.
Receive a full particle size distribution curve with D10, D30, D60, Cu, and Cc values. Export as CSV, AGS, or PDF report — ready for your engineering documentation.
To revolutionize soil analysis by providing a user-friendly, mobile, and affordable solution that increases both efficiency and accuracy in geotechnical engineering, construction industry, and related fields. We aim to contribute to the digitization and optimization of workflows through modern AI technologies while improving accessibility to precise analysis worldwide.
To become a leading provider of AI technologies in geotechnical engineering. Beyond developing existing products, we aim to create new applications and services covering various areas of geotechnical engineering—from soil analysis to construction monitoring and infrastructure planning. Through close collaboration with partners and customers worldwide, we strive to make a sustainable contribution to the modernization and automation of the industry.
Upload a soil sample image to analyze particle size distribution instantly.
Register for free to try grai4. For unlimited access, request an API key.
Auto uses EXIF data and a calibration database. Enter a value manually if you know your device's pixels-per-mm.
Drag & drop or browse
or download a sample image to test
Analyzing particle size distribution...
grai needs your phone's pixel density (PPM) for accurate scale detection. If your device isn't in our database or the default 70° angle-of-view approximation isn't precise enough, use our calibration tool below.
Watch the complete workflow — from data preparation to model output — and see how easy it is to predict particle size distribution with grai.
deepsoil's flagship product, grai, brings laboratory-grade precision to the field. No more waiting days for lab results.
Take a picture of the soil sample and get results in seconds using our proprietary computer vision algorithms.
Designed for the field. No specialized hardware required—just your mobile device and the grai platform.
Our AI model has been trained on thousands of samples to ensure accurate particle size distribution (PSD) mapping.
grai utilizes state-of-the-art deep learning architectures to segment and measure individual particles, accounting for lighting, scale, and texture variations.
Start free, upgrade when you need more. No hidden fees.
The standardised dark chamber for accurate image capture
Buy Boxes + LED Strip
~€30
Drill 5 cm hole, attach LED strip, fill bottom box with soil.
Our AI-powered solution offers significant advantages over traditional methods and competitors.
Rapid, cost-effective particle size analysis that replaces traditional sieve analysis and lab equipment.
Rapid assessment of soil stability and composition for construction projects. Required for infrastructure projects every 200m of soil investigation.
Optimize crop yield by understanding soil composition instantly at the farm site.
Monitor erosion and sediment changes in real-time across vast geographic areas.
Quick assessment of landslide susceptibility and slope stability in mountainous regions like the Alps.
Universities and research institutions can leverage AI-powered tools for teaching and advanced research.
deepsoil is developing a comprehensive suite of AI-powered geotechnical solutions. grai is our first product, with more innovations coming soon.
GRADING AI
AI-powered prediction of soil particle size distribution from smartphone images. Replace time-consuming laboratory procedures with a mobile, cost-effective alternative.
MOISTURE NET
AI solution for determining gravimetric water content of soils using the same photobox as grai. Provides comprehensive analysis of both particle distribution and moisture content.
LIQUID LIMIT NET
AI-powered solution for conducting Atterberg tests, specifically for determining the liquid limit. Automatically measures groove length through AI segmentation and controls the test automatically.
SOUND NET
AI-powered audio solution for determining soil particle size distribution. Analyzes the sound generated when soil material is poured onto the photobox to predict grain distribution.
Leading companies have expressed interest in piloting and testing deepsoil technology.
Leading provider of geotechnical analysis and design software. Interested in integrating grai AI capabilities.
Major geotechnical engineering consultancy exploring AI-powered soil analysis for their projects.
Specialized geotechnical company interested in field-ready analysis solutions.
Upper Austrian soil and building materials testing laboratory evaluating grai for quality assurance.
Global leader in geotechnical services, testing grai for rapid field soil assessment and quick site characterization.
Join these industry leaders and transform your soil analysis workflow.
Partnerships that power deepsoil's innovation.
The spin-off is based on innovative AI technology originally developed at the University of Natural Resources and Life Sciences, Vienna (BOKU).
Research institutions partnering with deepsoil to advance geotechnical AI — contributing data and receiving free grai access in return.
Using grai to cross-check and boost their sieve and sedimentation analyses.
Interested in an academic collaboration? Contact us.
The people behind deepsoil's innovative soil analysis technology.
Ph.D. in Environmental Engineering, BOKU Vienna
MBA in Project Management, WU Vienna
Passionate about bringing AI-powered solutions to geotechnical engineering and soil analysis. Enrico leads deepsoil's mission to revolutionize how we understand the ground beneath our feet.
Specialized in Machine Learning applications in Geotechnics during his Habilitation at BOKU. Experience leading tunnel construction projects.
A quick-start guide on how to use grai.
The original grai publication.
The dataset used in the grai1 publication.
The grai2 publication using a refined CNN architecture.
The latest grai3 publication (not peer-reviewed yet).
Join the professionals using deepsoil technology to change how we see the earth.
Special pricing available for universities and research institutions.
This short tutorial walks you through every step of using grai: calibrate your phone camera, prepare the photobox, place and photograph the soil sample, then upload the image to our platform for instant AI-powered particle size distribution predictions. From field to result — in minutes, not days.