Proprietary evidence stays confidential.

PoolTrack is a field-testing and movement-analysis application developed by Goodfield Studio in South Australia. It is intended to help authorised testers review how robotic pool cleaners move through a pool and return structured evidence to engineering or manufacturing teams.

Depending on the features used, PoolTrack may handle video recordings, optional pool photographs, coded test and unit names, firmware or build identifiers, tester notes, analysis settings, detection boxes, routes, movement candidates, confidence information and generated reports.

If you use the website contact form, it handles your name, reply email, selected topic, message and any optional organisation or coded test ID you provide. Testers should use internal codes rather than company names, product names or proprietary details wherever practical.

The current development build is designed to run within a controlled local workspace. Uploaded or imported recordings, pool photos, analysis databases, trained model files and reports are stored on the computer or server configured to run PoolTrack.

Local model analysis can process footage without sending frames to a hosted inference provider. This is the preferred mode for highly confidential or unreleased products. The person or organisation operating that computer remains responsible for its user access, backups, retention, device encryption and physical security.

Hosted processing is off unless an authorised user deliberately selects it and confirms that the material may be sent. When enabled, selected images and the minimum required project metadata may be sent to a third-party machine-learning provider. The development workflow currently supports Roboflow for optional annotation, inference and training. Full videos are not automatically uploaded to that service by PoolTrack’s local training workflow.

Third-party processing is subject to the provider’s terms, privacy practices, account settings and data location. It must be approved by the organisation that owns or controls the material. Local processing should be used when external disclosure is not authorised.

An optional photograph of the pool may be used to estimate its boundary and shape. Reports can present a schematic pool, coded unit and generic animated cleaner so the default view does not reveal product styling. This reduces exposure but does not anonymise the restricted original footage.

Goodfield Studio does not publish, sell, license, disclose or otherwise release a customer’s non-public footage, product design, specifications, model details, test results or reports to other companies, advertising networks, marketing channels, public datasets or public repositories.

Customer material is used only for the authorised test, support or development purpose. It is not used to train a shared or public model, demonstrate PoolTrack, develop a model for another customer, or create marketing material without the owner’s prior written authorisation.

Disclosure is limited to an authorised user’s direction, an approved service needed to perform the requested work, or a legal obligation. Goodfield Studio will use reasonable care to limit any approved disclosure to the minimum information required.

Local test material remains until an authorised user deletes it, removes the workspace or applies an organisation’s retention process. The interface’s deletion control removes the selected PoolTrack database record, its workspace copy of uploaded footage, optional pool photo, maps, reports and related analysis records. It does not remove an original file stored elsewhere.

Copies already exported, backed up or sent to another organisation must be managed separately. Data stored with an optional third-party provider must be deleted through that provider or its account controls.

The local pilot binds its dashboard to the host computer, rejects unrecognised hosts and cross-origin requests, uses per-process request protection and keeps credentials outside the repository. Production server deployments require organisation authentication, encrypted transport, least-privilege access and private storage appropriate to the test programme.

These controls are intended to prevent unauthorised access and data leakage. No responsible technology provider can promise that a security incident is impossible. Goodfield Studio will investigate a reported or detected incident, contain it and notify affected organisations when required by law or an applicable agreement.

The PoolTrack product website may receive standard web-request information such as IP address, browser type, requested page and approximate traffic statistics through Cloudflare’s hosting and security services. The site does not contain advertising trackers or ask you to create an account.

The contact form uses Cloudflare Turnstile to distinguish legitimate submissions from automated spam and Resend to deliver the enquiry to hello@goodfield.studio. Goodfield Studio also stores a limited enquiry record containing the category, reply email, message, creation time and a one-way hash of the submitting IP address in Cloudflare D1 for reporting and recovery. The raw IP address is not stored in that enquiry record.

You can use local analysis, choose whether to add a pool photo, decide whether hosted machine-learning services are enabled, delete local sessions and control which outputs are exported or shared. An organisation may apply additional policies to managed devices and test material.

PoolTrack is under active development. We may update this policy as hosting, account features, service providers or release capabilities change. The revision date above will show when the policy was last updated.

Questions, privacy requests or concerns can be sent to hello@goodfield.studio.

Goodfield Studio
South Australia, Australia