How we work
Our editorial standard, including what we refuse to publish
Quote or drop, never invent
Every clinical figure on this site comes from a named authority and is published in that authority's own words wherever the wording matters. Where a source states nothing, we publish nothing — we do not estimate, interpolate, or fill a gap with something that sounds right.
The clearest example is the warning-signs page: all fifteen signs are the CDC's exact text, including its nesting, because flattening a bullet there would change what the CDC actually said.
What we refuse to publish
The refusals are the part of this method that matters most, so they are worth stating plainly.
- No US childhood vaccine schedule from WHO's country registry — it lists Ebola, anthrax and yellow fever beside DTaP, with no field marking which are routine.
- No newborn triage list assembled from mismatched sources, because the authoritative US source is copyrighted and the open alternative gives UK emergency numbers.
- No iron percentage on Indian food pages, because ICMR-NIN publishes no pregnancy figure for iron and inventing one would be a clinical fabrication.
- No ranked "foods highest in" lists, which put baking powder above dairy for calcium and recommended liver — advised against in pregnancy — to pregnant readers.
We check sources against reality
Vendored data is verified against independently known values before it is published, and the checks are written as tests so they cannot silently drift.
This is not theoretical. The Indian food tables declare no units anywhere; read naively they overstated every calorie by 4.184× because energy is stored in kilojoules. The CDC schedule is an HTML table with merged cells; read naively it placed the MMR vaccine at 2 months instead of 12–15 months. Both errors look entirely plausible on the page. Both are now pinned by tests.
We publish our own weaknesses
The Sources page lists every dataset behind the site with its edition, when we retrieved it, what it powers, and what it gets wrong. That includes the uncomfortable entries: our baby-name corpus is the least-sourced content here, and it says so.
A reference work that only tells you its strengths is asking to be trusted rather than checked.
Calm is a design constraint
Health content is easy to write frighteningly and hard to write calmly without softening anything. Our rule is that a page must never leave a reader more alarmed than the evidence warrants.
So triage pages lead with the checklist rather than the alarm, reassurance sits in the same block as the warning, and where most people who check will find nothing, the page says so. No sign is softened to achieve that.
Guidance follows your country
Nutrient targets, vaccine schedules and health-system guidance differ by country, and presenting one country's figures as universal is a quiet way of being wrong. Readers in India see ICMR-NIN and the Universal Immunisation Programme; readers elsewhere see NIH and the CDC schedule. Every percentage names the authority it was calculated against.
Advertising and independence
Advertising pays for the site, and it never decides what we write, recommend, or leave out. No article is sponsored, and no advertiser sees a draft. We keep ads off urgent help; the kick counter, contraction timer, symptom tracker, medicine tracker and mental-health check; account and sign-in pages; and our policy, sources and contact pages — nobody checking a warning sign, logging something urgent, signing in, or reading our policies should have to scroll past one.
Who reviews this, and who doesn't
No page on this site has been reviewed by a doctor, midwife or any other clinician — we say that plainly rather than imply otherwise. Instead, every clinical figure is quoted or paraphrased from a named primary source (WHO, the CDC, ACOG, NIH, ICMR-NIN, the USDA and similar), cited on the page it appears on, so you can check it yourself without having to trust our judgment.
Swet Nisha, the site's founder, personally reads a growing set of pages as a parent — not a clinician — and checks that they read clearly and match what she found when she needed the same information herself. A page she has read this way says so, with her name, "parent reviewer," and the date, right next to that page's sources. Most pages do not carry this yet; the badge only ever appears where it is genuinely true.
If something reads as advice when it should read as reference, or looks wrong, out of date, or unclear, tell us through /contact and we will fix it or take it down rather than leave it standing.
The S2 "is it safe in pregnancy" food pages, and the vendored datasets behind /guide, are re-checked against their live sources on a monthly cadence (docs/CONTENT.md); nothing republishes automatically; a changed source is re-verified by hand before the page changes.
What we still need
Newborn-specific warning signs are the largest gap, and they are blocked rather than forgotten: the authoritative US source is copyrighted, and we will not write clinical thresholds ourselves. Closing it needs a licence or a clinician's authorship, and until then the page says plainly that we publish no such list.
Last reviewed 27 September 2026.