Every score on Everyday AI Radar is calculated, not felt. We read the makers' official spec sheets, stores and pricing pages, record each figure with the date we read it, and apply the same formula to every product of a category. The current data was checked on Oct 29, 2026.
What we do not do
- We do not test products in a lab or in the field, and we never write as if we had. Battery life, language counts and recording times are the makers' figures.
- We do not measure translation accuracy, transcription quality, video quality, voice realism, comfort or app reliability. These matter, and our scores do not capture them.
- We do not assess medical performance. For baby monitors, we only repeat what each maker states about breathing alerts and FDA status, word for word where it matters.
- We do not use user star ratings, and we do not estimate missing figures.
The formula
- Points per criterion (0 to 10). Numeric criteria are placed between the lowest and the highest value of the category: the worst value gets 3 points, the best 10, linearly (or on a logarithmic scale for prices, battery days and language counts). If every product has the same value, each gets 7. Yes/no and level criteria get fixed points, listed below.
- Raw score. The weighted average of the points of the criteria that are known for the product. A missing criterion is removed and the remaining weights are renormalized: it never counts as zero.
- Who gets a score. A product needs a published price, at least 2 known criteria besides the price, and known criteria worth at least 67% of the non-price weight of its category (criteria that no product of the category publishes are left out of that total). Otherwise it is listed without a score or rank, with the reason shown.
- Displayed score. Displayed score = 7.8 + 1.7 × raw score ÷ 10, rounded to one decimal, between 7.8 and 9.5. The scale is editorial: it only spreads products that already passed our data bar. Rankings use the unrounded raw score, then the lower price.
- Prices. The current price on the official store, using the maker's displayed promotion when there is one (the list price is shown next to it). For apps, the cheapest paid plan per month on monthly billing, or on annual billing when that is the only price shown.
Criteria and weights by category
AI translator earbuds
| Criterion | Weight | Points |
|---|
| Offline languages | 3 (30%) | linear, category min to max: higher is better |
|---|
| Online languages | 1.5 (15%) | linear, category min to max: higher is better |
|---|
| Battery per charge | 2 (20%) | linear, category min to max: higher is better |
|---|
| No subscription needed | 1.5 (15%) | fixed points: none = 10, partly = 5, required = 0 |
|---|
| Price | 2 (20%) | logarithmic, category min to max: lower is better |
|---|
Handheld translators
| Criterion | Weight | Points |
|---|
| Offline languages | 3 (29%) | linear, category min to max: higher is better |
|---|
| Online languages | 2 (19%) | linear, category min to max: higher is better |
|---|
| Mobile data included | 1.5 (14%) | linear, category min to max: higher is better |
|---|
| Screen size | 1 (10%) | linear, category min to max: higher is better |
|---|
| No subscription needed | 1 (10%) | fixed points: none = 10, partly = 5, required = 0 |
|---|
| Price | 2 (19%) | logarithmic, category min to max: lower is better |
|---|
AI voice recorders
| Criterion | Weight | Points |
|---|
| Free transcription minutes / month | 2.5 (28%) | linear, category min to max: higher is better |
|---|
| Paid plan per year | 1.5 (17%) | linear, category min to max: lower is better |
|---|
| Recording time per charge | 2 (22%) | linear, category min to max: higher is better |
|---|
| Weight | 1 (11%) | linear, category min to max: lower is better |
|---|
| Price | 2 (22%) | logarithmic, category min to max: lower is better |
|---|
AI glasses
| Criterion | Weight | Points |
|---|
| Battery per charge | 2 (21%) | linear, category min to max: higher is better |
|---|
| Camera | 1.5 (16%) | linear, category min to max: higher is better |
|---|
| Built-in display | 2 (21%) | fixed points: yes = 10, no = 4 |
|---|
| Prescription lenses | 1.5 (16%) | fixed points: yes = 10, no = 2 |
|---|
| Price | 2.5 (26%) | logarithmic, category min to max: lower is better |
|---|
Smart baby monitors
| Criterion | Weight | Points |
|---|
| Breathing-motion alerts (maker claim) | 2.5 (23%) | fixed points: yes = 10, no = 3 |
|---|
| FDA status as stated by the maker | 2 (18%) | fixed points: yes = 10, no = 4 |
|---|
| Video resolution | 1.5 (14%) | linear, category min to max: higher is better |
|---|
| No subscription needed | 2 (18%) | fixed points: none = 10, partly = 5, required = 0 |
|---|
| Local video storage | 1 (9%) | fixed points: yes = 10, no = 3 |
|---|
| Price | 2 (18%) | logarithmic, category min to max: lower is better |
|---|
GPS dog collars
| Criterion | Weight | Points |
|---|
| Live GPS over cellular | 1.5 (16%) | fixed points: included = 10, optional = 6 |
|---|
| Required plan per month | 2.5 (26%) | linear, category min to max: lower is better |
|---|
| Battery life | 2 (21%) | logarithmic, category min to max: higher is better |
|---|
| Virtual fence | 1 (11%) | fixed points: yes = 10, no = 2 |
|---|
| Price | 2.5 (26%) | logarithmic, category min to max: lower is better |
|---|
AI video generators
| Criterion | Weight | Points |
|---|
| Free plan | 2 (21%) | fixed points: yes = 10, no = 0 |
|---|
| Entry price per month | 2.5 (26%) | logarithmic, category min to max: lower is better |
|---|
| Languages | 1.5 (16%) | logarithmic, category min to max: higher is better |
|---|
| Voice cloning | 1.5 (16%) | fixed points: yes = 10, enterprise = 5, no = 0 |
|---|
| Max export resolution | 1.5 (16%) | linear, category min to max: higher is better |
|---|
| No watermark on free plan | 0.5 (5%) | fixed points: yes = 4, no = 10 |
|---|
AI voice generators
| Criterion | Weight | Points |
|---|
| Free plan | 2 (24%) | fixed points: yes = 10, no = 0 |
|---|
| Entry price per month | 2.5 (29%) | logarithmic, category min to max: lower is better |
|---|
| Languages | 2 (24%) | logarithmic, category min to max: higher is better |
|---|
| Voice cloning | 2 (24%) | fixed points: yes = 10, enterprise = 5, no = 0 |
|---|
Editor decisions
Some figures need a rule. When a maker states that offline packs are not included and publishes no offline language count, we count zero offline languages. When a maker's pages contradict each other (for example on whether a membership is required), we use the less favorable reading and say so on the product page. When a maker's product image is a banner or carries promotional badges, we show the product name on a plain card instead. Every correction is logged in our data files with its reason.
Partners and commissions
Today no link on this site is an affiliate link. If we join affiliate programs, partners may rank higher in editorial blocks such as "Our pick" or "Best for"; any such placement will be disclosed next to the block. The scores and the order of the rankings will keep coming from the formula above, which ignores commissions. See our affiliate disclosure.
Corrections
Found a wrong figure? Write to us with the official source at our contact page. We correct data, recalculate scores and update the check date.