How SparkyStudy handles course material

The product rules we hold ourselves to, and the specific commitments we make to publishers, trade schools, and standards bodies.

The core rule

SparkyStudy does not ship, redistribute, or embed copyrighted curriculum. Study content comes only from material each student uploads themselves, and stays on that student's account. Nothing derived from one account's uploads is shown to another.

This rule is written into the codebase and enforced in the architecture, not just the policy. If you're a publisher reviewing SparkyStudy for a potential collaboration, the three commitments below are the ones that matter.

Three commitments, in plain words

1. Student uploads only

SparkyStudy ships with no built-in question banks, flashcards, or study guides derived from licensed curriculum. The seed flashcards included are self-written basics (Ohm's Law and foundational trade theory). Everything beyond that is generated at the moment a student uploads a document they own a copy of — and only that student sees it.

2. Per-account scope

A student's uploaded modules, generated flashcards, and generated exam questions live on that student's account only. They are not visible to other students, not aggregated into a shared library, and not surfaced in search. If student A uploads their copy of a textbook chapter, student B cannot see it, generate from it, or benefit from its upload. If student A deletes their account, the derived content is deleted too.

3. Cite, never reproduce — for codes and standards

Where the Canadian Electrical Code, the US National Electrical Code, OSHA, NFPA 70E, CSA Z462, or similar standards are referenced, SparkyStudy quotes section numbers, rule numbers, and table numbers only — never the code's wording. Section numbers are facts (Feist Publications v. Rural Telephone Service). The code text itself is copyright and licensing-controlled. We don't ship it, don't paraphrase it, and don't let the AI tutor quote it back to students.

How these show up in the product

Uploads are detected and classified. When a student uploads a document, the backend classifies it as either student notes or publisher material. Material flagged as publisher content is treated accordingly — generated content is grounded to that student's uploaded text, not reproduced from it, and never shared across accounts.
Generated questions are grounded. The question-generation prompts require every question to be anchored to a specific sentence in the student's source text. Hallucinated or outside-sourced questions are refused — the model returns an empty array instead of guessing.
The AI tutor cites, doesn't quote. The tutor is instructed to reference code sections by number, say "I don't know" when uncertain, and self-rate confidence on every response (HIGH / MEDIUM / LOW / UNKNOWN) so students know when to verify in the code book.
Code-table calculators are gated on licensing. Where a calculator would need to embed values from a copyrighted table (CEC Table 2, NEC Table 310.16, etc.), it stays offline until written permission exists. Current status: eight calculators are scoped NEC-only (US-published tables with explicit citation); three remain under review pending CSA permission for CEC-sourced values.

If you're a publisher or standards body

If your organization publishes training material or standards our users may upload, and you want to discuss a collaboration — permission to reference, co-authored content, revenue-share on students who buy from you through us, or anything else — we'd like to talk.

Contact: info@sparkystudy.com. We can walk through the architecture, the content-flow, and anything else you'd want to see in detail.