Trust Center
Security and privacy, in plain terms
How Skill Reality handles the images, records and accounts from your floor. Anything we have not published yet is available on request.
One frame kept, faces blurred
No continuous video. Retention set by you.
Verified · 08:58:41
At a glance
Six things to know
One image at critical steps
The glasses camera captures a single image only at the steps you mark critical. No continuous video is recorded.
Faces blurred on the device
Faces are blurred on the device, before the image is uploaded. Kept images stay blurred.
Retention set by you
You choose how long verification images and records are kept.
Encrypted in transit and at rest
Procedure data and step logs are encrypted in transit and at rest.
Built on Microsoft Azure
The Skill Reality platform runs on Microsoft Azure in Canada.
Single sign-on
Sign in with Microsoft Entra ID, Google Workspace or Okta.
How data flows
From the glasses to your dashboard
What we collect and why
Only what the record needs
| Data | Why | How long it is kept |
|---|---|---|
| Verification image | To check that a step you marked critical was done right. | Blurred on the device before upload; kept for as long as you set. |
| Step record: worker, procedure, step, time and result | Your audit, quality and training record. | For as long as you set. |
| Account: name and work email, from your sign-in provider | To sign workers in and record who did each step. | For the life of the account, plus 30 days. |
| SOPs and procedures you upload to Forge | To build the guided procedures. | Until you delete them. Deleted with the account. |
| Remote expert video | Live help from an expert who sees what the worker sees. | Not recorded. Sessions are live only; nothing is stored. |
What we do not collect
- No continuous video from the glasses.
- No audio recording for verification.
- No images outside the steps you mark critical.
Camera and worker privacy
For the people wearing the glasses
The camera captures one image only at the steps your team marks critical. Faces are blurred on the device before the image is uploaded, and there is no continuous video. We wrote a plain-language page for workers, their union and HR, with a one-page sheet for the break-room board.
Infrastructure and access
Where it runs and who can see it
- Hosting
- Microsoft Azure
- Azure region
- Canada
- Encryption
- Procedure data and step logs are encrypted in transit and at rest.
- Encryption detail (protocols, key management)
- TLS 1.2 or higher in transit. AES-256 at rest, with Microsoft-managed keys.
- Access control
- Role-based: workers and managers see only what is relevant to their role.
- Single sign-on
- Microsoft Entra ID, Google Workspace, Okta
- Deletion timeline after a retention period ends or a contract ends
- Data is deleted within 30 days after the retention period ends or the contract ends. Backups are purged within 90 days.
Subprocessors
Companies that process data for us
| Subprocessor | Purpose | Used for | Data location |
|---|---|---|---|
| Microsoft Azure | Hosting, processing and storage for the Skill Reality platform | Platform | Canada |
| LiveKit | Live video for remote expert sessions | Platform | LiveKit Cloud, North America |
| OpenAI | AI model provider (ChatGPT models) | Platform | United States, under its standard API terms |
| Anthropic | AI model provider (Claude models) | Platform | United States, under its standard API terms |
| Vercel | Hosting for this website (skillreality.com) and its forms only; no customer plant data | Website | United States (Washington DC, iad1) |
| Resend | Email for this website's forms only; no customer plant data | Website | United States (us-east-1) |
Website subprocessors handle this website and its forms only, not data from your floor.
AI model providers
- What AI providers receive
- The text of your SOPs, for Forge to parse, and the single critical-step image, with faces already blurred on the device, for verification. Nothing else.
- Where AI processing happens
- In the United States. Data sent for AI processing leaves Canada.
- Training on your data
- OpenAI and Anthropic do not use data sent through their APIs to train their models by default, under their standard API terms.
- OpenAI: retention
- Up to 30 days in OpenAI's abuse monitoring logs, under its standard API terms.
- Anthropic: retention
- Available on request
Compliance roadmap
What is planned and what you can ask for
SOC 2
Planned Type I planned for Q4 2027. We complete security questionnaires today.
Independent penetration test
Planned
Data processing agreement (DPA)
In preparation Contact security@skillreality.com.
Report a vulnerability
Found a security issue?
Email security@skillreality.com with what you found and the steps to reproduce it. We read every report and reply.
Please give us reasonable time to fix an issue before you disclose it, and do not access data that is not yours or disrupt the service while testing. Our contact details are also in security.txt.
Downloads
