Last updated: March 12, 2026
Powmetrix is a powder flowability prediction tool. This policy explains what data we collect, how we use it, and your rights — in plain language.
When you create an account, we store your email address and a hashed password (we never store your actual password). We also store your organization name and your role within it (admin or member).
We do not collect your real name, phone number, address, or payment details.
When you create materials, we store the particle size distribution parameters, physical properties (density, surface energy), and optionally your measured flowability coefficient (FFC). This data is used to generate predictions.
Your material data belongs to your organization. No other organization can see, access, or query your materials. Every database query is scoped to your organization.
Prediction results (FFC values, confidence scores) are computed on-the-fly and cached temporarily in server memory. They are not permanently stored in the database.
We keep standard web server logs (IP address, pages visited, timestamps) for up to 14 days for debugging and security monitoring. These are never shared with third parties.
Powmetrix includes an optional program where you can contribute anonymized material data to help improve the prediction model for all users. Here's how it works:
No data is ever shared unless you explicitly toggle "Contribute to model improvement" on a specific material. The default is off. You must also have a measured FFC value for that material — without it, the data cannot be used for training.
Only numerical physics parameters: particle size percentiles, bimodal fit parameters, density, surface energy, and your measured FFC value. That's 12 numbers.
Contributed data is fully anonymized. After stripping identifiers, it is not possible to trace a data point back to any specific organization or user.
Contributed data is combined with our existing training dataset to retrain the prediction model. A new model is only deployed if it measurably outperforms the current one. Contributed data is never displayed to other users, exposed through any API, or shared with any third party.
Turn off the "Contribute" toggle on any material, and it will be excluded from the next model update. Note that models already trained on your data cannot be retroactively modified, but future updates will not include your withdrawn data.
| Protection | Details |
|---|---|
| Encryption in transit | All traffic encrypted via HTTPS (TLS 1.2+) through Cloudflare |
| Password security | Passwords hashed with bcrypt (never stored in plaintext) |
| Session security | HttpOnly cookies, CSRF protection, short-lived tokens (15 min) |
| Rate limiting | Login and registration attempts are rate-limited to prevent abuse |
| Tenant isolation | All queries are scoped to your organization — no cross-org data access |
| Backups | Daily encrypted database backups, 14-day retention |
We use a minimal set of external services:
| Service | Purpose | Data involved |
|---|---|---|
| Cloudflare | DNS, CDN, DDoS protection, SSL | IP addresses (subject to Cloudflare's privacy policy) |
| Hostinger | Server hosting | All server data is stored on their infrastructure |
| Google Fonts | Web fonts | Standard font loading request (no cookies) |
We do not use Google Analytics, tracking pixels, advertising networks, third-party AI/ML APIs, or any service that processes your material data. All predictions run entirely on our own server.
| Data | Kept for |
|---|---|
| Account & material data | Until you delete it or request account deletion |
| Prediction results | Temporary (in-memory only, cleared on server restart) |
| Server logs | 14 days (automatic rotation) |
| Database backups | 14 days (automatic rotation) |
We may update this policy as Powmetrix evolves. Significant changes — especially regarding the Data Contribution Program — will be communicated through the application.
For privacy questions, data deletion requests, or concerns:
Email: [email protected]