Rules, not guesses.

Matching is deterministic

Every program in our database carries structured eligibility rules, province, age range, income thresholds, household circumstances, business size, and required characteristics. Your answers are evaluated against those rules directly. The same answers always produce the same results.

No language model decides whether you qualify. That's a deliberate constraint: a model that's wrong about eligibility is worse than useless, because it's wrong persuasively.

Strong versus possible

A strong match means every eligibility rule we could check passed against an answer you actually gave.

A possible match means the program looks relevant but something is unresolved, usually an optional question you skipped, or a condition only the administrator can confirm. If you skip the income question, an income-tested program can never be a strong match, no matter how well everything else lines up. We would rather under-promise.

Amounts come from the source or not at all

If an official page states a figure, we show it and say where it came from. If it doesn't, we show “not stated by source” rather than estimating. You will never see an invented savings total on this site, because a made-up number is the fastest way to make a real service look like a scam.

Freshness is the hard part

Program rules change constantly. Every listing shows the date a human last checked it against its official page, and programs carry an explicit status, active, closing soon, temporarily closed, expired, or needs verification. A paused program stays visible and is labelled as paused, because knowing a rebate isn't available right now is worth as much as knowing one is.

On top of the human checks, we re-fetch every program's official page daily and look for the things that signal a change: a closure notice, an intake window ending, a reworded requirement, or a dead link. Anything that trips gets flagged for review.

That automation is deliberately one-directional. It can mark a program closed or in need of re-verification, but it can never mark one active or stamp it as verified. Only a person does that, after opening the official page. An automated system that could promote a program would eventually promote a wrong one.