Designed for school use
Hypersketch is a teacher-directed visual creative-learning environment, not an open-ended general-purpose AI chatbot. Students use AI to explore and develop ideas while remaining responsible for what is meaningful, what is worth pursuing and what they create.
The safeguards described below are technical, pedagogical and operational — not simply a content filter.
This page is written for school leadership, teachers, IT and privacy reviewers. The Application Privacy Policy is the authoritative statement on data handling.
Hypersketch is a bounded visual creative-learning environment, used within teacher-directed learning. It is designed around image generation and editing for creative tasks, structured so that students continue making decisions, and intended to support ideation rather than replace student ownership.
It is not any of the following:
There is no unrestricted general-purpose chatbot and no open web browsing. Students principally use AI for bounded image generation and editing within a creative task.
Prompts are screened for inappropriate content, including profanity, sexual content, obscenity and other prohibited material.
Generated images are screened before being shown to the student. This is a second layer, applied after input screening.
Activities take place within structured learning. For facilitated younger-student workshops, Hypersketch facilitators work alongside school teachers. Classroom overview functionality can allow teachers to see what students are working on in near real time.
Hypersketch is deliberately designed so that good outcomes do not come from simply asking AI for a finished answer. The creative workflow requires students to notice, judge, connect, reframe, choose and direct repeatedly as an idea develops.
notice → judge → connect → reframe → choose → direct
This productive friction is intentional. It makes students exercise the mental moves involved in creative thinking rather than bypass them. If a student tries to jump straight from a generic instruction to a final result, the outcome is often weak, incoherent or unresolved; the quality emerges through the student’s interpretation and direction.
This is therefore not only a learning objective layered on top of the software. It is built into how Hypersketch works. The system generates possibilities, but human-directed outcomes emerge through repeated decisions made by the student.
The human is not brought back in at the end. The human is in the loop throughout.
AI can create a visual surprise. The student still has to decide why it matters.
For younger learners, AI interaction can be further mediated or restricted according to school requirements. We are willing to work within additional school AI-use, supervision and privacy requirements.
Student prompts, sketches, uploads, project work and generated images are not used by Hypersketch to train, retrain or fine-tune AI models.
Hypersketch also does not opt student content into third-party model training.
Students authenticate through a school-authorised email account. We seek to minimise unnecessary identifying information and use pseudonymous internal identifiers where practicable. Students are not required to provide details such as a home address, telephone number or date of birth.
Hypersketch does not:
Saved Web App projects are currently stored in Hypersketch-managed cloud infrastructure in Singapore. Saved project work is distinct from the temporary copies created while an AI generation is being processed.
Some AI functions require temporary processing by specialist cloud AI infrastructure outside Singapore. The processing location depends on the function and the service provider. Only the content necessary for the requested AI action is sent to the relevant service.
Depending on the function and provider, temporary processing may occur in:
Infrastructure locations can change. Temporary processing data generally exists only while the computing environment is carrying out the request, and is removed when that environment shuts down — normally within approximately one hour for the relevant serverless workflows.
Certain OpenAI API functions may retain limited request and response information in abuse and security monitoring logs for up to 30 days, in accordance with the applicable API policies.
Full detail is set out in the Application Privacy Policy.
Student workspaces are private from other students by default. Other students see work only when the student deliberately shares, presents, exports or uses an approved classroom sharing workflow.
Class Stream is one such optional sharing workflow, used for critique, presentation, collaboration or teacher review. It is a deliberate act by the student, not a default broadcast of student work. Teacher and administrator visibility can be configured for the school’s deployment.
For deployments that require parent or carer communication, Hypersketch can provide school-facing privacy and consent information covering student sign-in, saved work, AI processing, data use and consent.
If your IT, privacy or leadership team needs to review Hypersketch before a workshop, pilot or deployment, these are the relevant documents. Questions can be sent to support@hypersketch.com.