Spatial Intelligence & User Anonymity are not mutually exclusive
We are an EU-based privacy-tech company providing the Privacy and Location Reasoning Layers for the physical world. We enable organizations to harness the power of location intelligence without the radioactive risk of tracking individuals.
Our philosophy is simple: Protect privacy and keep the insights.
What we solve:
- GDPR Compliance: Automatically protect telemetry and private addresses data at the point of ingestion to meet the world’s strictest privacy standards.
- Secure Data Liquidity: Eliminate the risk of PII leaks when sharing spatial datasets with partners, subcontractors, or the public.
- LLM & Chatbot Privacy: Prevent sensitive addresses and precise coordinates from “leaking” into AI models and chat interfaces.
- High-Performance AI/ML: Replace messy, high-cardinality address data with optimized spatial features that improve model accuracy across entire continents.
- Socio-Economic Enrichment: Access high-resolution datasets without compromising the privacy of the populations they describe.
Under the hood
We provide three baseline services:
- Batch observations obfuscation: transforms given set of coordinates and observations (values) at those locations into the statistically similar dataset. You protect not only spatial coordinates, but also exact values of measurements – that could be signals from IoT or transactions value. Algorithm is based on the strong foundations, obfuscated data statistical properties are similar to the unsecured dataset in 95-99% for almost all scenarios. This dataset can be shared with other parties or event publicly.
- Chatbot protection – location insights: as companies integrate AI, precise location data often “leaks” into model prompts and history. Our Contextual PII Analyzer acts as a real-time firewall for conversational interfaces.
- Boosting AI/ML models with location features: we enhance your models’ accuracy analyzing spatial relations in training dataset or in inference responses. Moreover, we can point you to locations that are undersampled showing predictions uncertainty levels.
Early Access
We offer two early-access modules that you can start using after joining our waitlist (and being selected as a trusted partner).
- The Camouflage API (Ca-W360L-SEC) – obfuscate your static coordinates and values before sharing with others. Get insights about undersampled regions.
- The Saturation API (Sa-W360L-AI) – enrich your location data with socio-economic variables, create new features for your AI/ML models.
Ethics
- No snooping around – we do not store your data in any form, raw or processed, including weights in AI/ML-models.
- No tracking individuals – our tools may not be used to track individuals, and those capabilities are suppressed. You won’t be able to track anyone using their personal data.
- No De-anonymization. We do not provide services capable of deanonimization of individuals. We are not operating from Mordor.
- Close cooperation with security researchers – our tools and codebases are open for security researchers, especially from EU. Unfortunately, we won’t publish open source code for everyone.
Partnerships
- CODE:ME Foundation: Cybersecurity Events and Courses, HQ in Poland 🇵🇱 🇪🇺
Join the Waitlist!
Partnering for Good: Our Ethical Commitment
We don’t just protect data; we protect people. Because our technology is designed to shield individuals from invasive tracking, we have a strict Ethical Alignment Policy. We are hand-selecting our first three partners based on their commitment to human rights and transparent data practices.
Our Vetting Process:
- Application: Tell us about your organization and your specific use case.
- Alignment Review: We perform due diligence to ensure our tools will not be used for surveillance, predatory enforcement, or any activity that compromises fundamental human freedoms.
- A dialogue: If our values align, we’ll meet to discuss how our Privacy Layers can strengthen your mission.
Contact
(You might be asked to confirm that you are a human, please do it before sending the message).