Pocket.Science
We make apps no one else can
Craftsmanship: scientific instruments disguised as apps. Rocket science under the hood: physics, sensor calibration, machine learning. The user never has to see any of it. Bring us a phone, a camera, a sensor, or a dataset and a question about the real world; in 2–6 weeks you hold a working app that measures, classifies, and explains.
By the team behind iSPEX (featured in Nature) and Black Hole Finder (Space.com, Gizmodo, The Verge). 15 years of turning smartphones into scientific instruments.
You want to observe, measure, or classify something, without hiring a technical team
Businesses with a quality, inspection, or monitoring problem. Agencies that need field data. Startups that need a sensing prototype before they hire engineers. Research teams that need it built, not proposed. You know what you need to understand about the real world; we build the instrument that captures it, classifies it, and puts the result on your screen.
Four ways in
Phone as inspector
A phone-based inspection or quality-control tool: point the camera, get a measurement or pass/fail classification. Field-ready.
Field observation + AI
An observation app with AI classification for ecology, water, sky, materials, or anything your team or the public observes outdoors.
Private AI lab box
A local AI machine that analyzes your images, audio, or sensor streams on-premises. Your data never leaves the building.
Something else
Bring us your measurement problem. If it can be sensed, photographed, or logged, we can probably make an instrument out of it.
From problem to prototype
1. Problem call
A free call where you describe what you want to measure or classify. We tell you honestly whether we can crack it.
2. Fixed quote
A scoped, fixed-price proposal within days, from €5,000 depending on scope. No hourly billing, no surprises.
3. The build
2–6 weeks of building: sensing, data pipeline, AI classification, interface. You see progress weekly.
4. Working prototype
You get a working instrument plus an optional support retainer. Our own instruments have run in production for 15 years.
Real science in real hands, from the North Pole to Zimbabwe
Proof, not promises
iSPEX turned a smartphone into a spectropolarimeter: featured in Nature, validated at Plymouth Marine Laboratory against equipment costing tens of thousands of euros. Black Hole Finder lets citizen classifications automatically trigger robotic telescopes at Las Cumbres Observatory: 890,000+ classifications, 1,100+ automated follow-ups, running unattended today.
Instruments that survive contact with reality
Field calibration, messy data, real users, long-term operation. The unglamorous parts are where most prototypes die. Ours haven't: our platforms have run continuously for over a decade with millions of users.
Why AI won't replace this
A code generator can write an app. It cannot calibrate a smartphone camera against a marine laboratory, model the physics of scattered light, or fuse telescope pipelines with human judgment. The hard part of our work is measurement and field reality. Software is just how we deliver it.
Apps that mobilize millions
Our first app, Transit of Venus in 2012, hit 100,000+ downloads in one week. Since then: millions of downloads, communities of contributors, and coverage by Space.com, Gizmodo, TechRepublic, and The Verge. We build the whole loop: instrument, experience, and adoption.
Frequently asked
Can you build an app that classifies images or sensor data with machine learning?
Yes, that is our core work. Black Hole Finder combines citizen classifications with ML models on live telescope data; we build the same kind of pipeline for inspection, ecology, water, materials, and quality control.
What does a custom measurement app cost?
Fixed price from EUR 5,000 depending on scope, quoted up front after a free problem call. No hourly billing. Optional support retainer after delivery.
Why not just generate an app with AI tools?
Code generation gets you software. It does not get you a calibrated instrument: sensor physics, validation against reference equipment, and field reality are the hard part. That is the part we have done for 15 years.
Do you work with our existing hardware or data?
Yes. Smartphones, cameras, microscopes, lab sensors, drones, or an existing dataset. If it can be sensed, photographed, or logged, we can usually build the app around it.
Describe your measurement problem
Two sentences is enough. We reply within 24 hours, usually with a few sharp questions and an honest assessment.
Prefer email? [email protected] · +31 45 20 31 008