Microgravity research, booked and analysed like a cloud service.
SpaceLab is a software platform that gives universities, startups and pharmaceutical companies access to microgravity experiments on existing flight carriers, and analyses the results with AI. We don't build or own hardware: we aggregate capacity from partner carriers and handle booking, experiment setup, monitoring and analysis in one place.
Origin: 3rd place at NASA Space Apps Challenge Valencia 2025.
Protein crystals grown in microgravity are often larger and better ordered, which helps drug design. But access is fragmented across providers, priced for big budgets, and results still need slow manual image review. Smaller institutions are left out.
- Pro — universities, startups and education. Lower price, flies on available capacity, and results become open science after an embargo period, so every Pro experiment adds to a shared public dataset.
- Premium — pharma and biotech. Priority on every mission and confidential results.
Premium revenue cross-subsidises Pro access. Rules are implemented in api/app/rules.py.
| Folder | What it is |
|---|---|
ml/ |
Crystal-detection model (EfficientNet on MARCO): training, evaluation, threshold tuning, TFLite export for edge hardware. Original work by Mihail Stankevich in ml/legacy/. |
api/ |
Platform backend: slot marketplace, experiments, telemetry, AI analysis on upload, public Open Science catalog. |
docs/ |
Product and application planning. |
Coming next: web/, the client app built from the Dashboard and Slot Marketplace designs.
# ML
cd ml && pip install -e ".[dev]" && pytest -q
# API
cd api && pip install -r requirements.txt && pytest -q
SPACELAB_ADMIN_KEY=dev uvicorn app.main:app --reload # docs at http://localhost:8000/docs- Mar Minguez — product, UX and data
- Mihail Stankevich — machine learning
To be decided by the team before making the repository public.