1. About these terms
These terms govern your use of the tallGuard website, customer account features, and tallGuard pond-monitoring software. By creating an account, pairing a device, or using these services, you agree to these terms.
2. The service
tallGuard is designed to help pond keepers monitor for herons using local camera processing on supported Raspberry Pi and Edge TPU hardware. Core detection, recording, heartbeat emails, and deterrent functions are designed to operate locally where configured. Account, notification, and other cloud features require internet access and may be unavailable from time to time.
3. Detection is an aid, not a guarantee
Herons, cameras, lighting, weather, network conditions, hardware, and model confidence can affect results. tallGuard cannot guarantee that every heron, fish-loss event, camera fault, or other event will be detected or prevented. You remain responsible for your pond, animals, equipment, and chosen physical deterrents.
4. Your account and device
Keep your account access and device hardware secure. Only pair devices that you own or are authorised to manage. Do not share account access, bypass ownership checks, misuse pairing codes, or attempt to access another customer's devices, activity, recordings, or account information.
5. Camera and lawful use
You are responsible for placing and operating your camera lawfully, including obtaining any permissions needed for property, people, or areas that may be captured. Do not use tallGuard for unlawful surveillance or to capture areas where people reasonably expect privacy.
6. Acceptable use
Do not interfere with the service, probe for vulnerabilities, reverse engineer cloud access controls, distribute device credentials, expose camera endpoints to the public internet, or use tallGuard in a way that harms people, wildlife, systems, or other customers. We may suspend access needed to protect the service, customers, or devices.
7. Model and third-party components
tallGuard's heron-detection model was developed using TensorFlow's SSD MobileNet v2 FPNLite object-detection architecture and the ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8 training configuration. It was fine-tuned using independently collected personal images of herons. Customer camera footage is not used to train or improve the model. Third-party components remain subject to their applicable licence terms and do not imply endorsement of tallGuard by TensorFlow, Google, or the COCO project.
8. Third-party services and links
tallGuard may link to or integrate with third-party services, including identity providers, cloud infrastructure, video hosts, affiliate retailers, and donation platforms. Their services are governed by their own terms and privacy practices. We are not responsible for third-party sites or transactions.
9. Changes and availability
We may update these terms, improve or change features, or discontinue features where reasonably necessary. We will publish material changes on this page with a revised update date. Continued use after a change takes effect means you accept the updated terms.
10. Liability
To the extent permitted by law, tallGuard is provided on an as-available basis and we are not liable for indirect or consequential loss, including loss of fish, property damage, lost data, or lost profits. Nothing in these terms excludes liability that cannot lawfully be excluded.