Introducing the Privacy Keeper Blog
Privacy Keeper is a Windows desktop application that automatically detects and blurs faces and license plates in 360° panoramic images. It runs entirely on your own hardware — no cloud upload, no telemetry, no subscription to a processing service. We built it for the people who capture equirectangular panoramas professionally: surveyors, mapping companies, 360° photographers, and organisations that need to publish street-level imagery without exposing personal data.
We have had a Guides section since launch — practical how-to content on blurring workflows and GDPR compliance. The blog is different. It is where we will write about the broader context: how to evaluate tools in this space, which cameras work well for mapping, how offline processing compares to cloud services, and what we are building next.
What Privacy Keeper does
The application processes equirectangular (2:1) panoramas up to 32K resolution. It uses AI-powered detection for faces and license plates, NVIDIA CUDA acceleration with automatic CPU fallback, and a manual review layer so you can confirm, correct, or add detections before export. Images never leave the machine during processing — the anonymisation step runs locally, and the exported files are the only output.
Performance on tested hardware: a 12,288 × 6,144 Mosaic 51 panorama takes approximately 9 seconds on an RTX 3090, or 26 seconds on CPU. An 8,192 × 4,096 Insta360 X4 frame takes approximately 4.3 seconds on GPU. Batch processing with checkpoint/resume handles large datasets without manual restarts.
What we will cover here
- Privacy & Compliance. What GDPR requires when you publish street-level imagery, and how to build anonymisation into your pipeline rather than treating it as an afterthought.
- 360° Workflows. How mapping teams structure their pipelines — from capture to anonymisation to delivery.
- Buying Guides. Which cameras produce equirectangular output that is practical to anonymise at scale, and how they compare on resolution, processing time, and total cost.
- Product updates. What we have shipped, what changed, and why.
Who this is for
If you capture 360° panoramas and need to anonymise them before publishing — for GDPR, for client requirements, or for your own data policy — this blog is written for you. We are not writing generic privacy explainers. The context is always equirectangular imagery, on-device processing, and the practical decisions that come with operating at scale.