The nine quotations we withdrew
This site publishes 18 milligram quotations — the seller’s own words for how much NAD+ a price buys, printed beside every figure on the cost-per-milligram board and on each seller’s review page. They are the evidence the whole site rests on: without them our division is an assertion.
Until 5 September 2026, 9 of those 18 were not quotations. They were reconstructions — near enough to be plausible, different enough to be untrue. A label and its value from two separate rows joined into one sentence. An asterisked footnote folded into a parenthesis. Spacing closed up. Clauses re-ordered. Each was rendered inside typographic quote marks and attributed as the seller’s own words.
And every milligram figure they carried was right. Not one number on this site changed when they were corrected. That is the part worth sitting with: a price audit checks that the arithmetic reconciles, and this defect reconciles perfectly, forever.
What went wrong, and why every number was right
A quotation is a promise that the words between the marks are the words on the page. The nine below broke that promise in small ways, and the smallness is the danger: nobody reading them would have suspected anything, because the meaning was preserved and the numbers were exact.
The reason it happened is mundane. Sellers do not print a milligram total as one tidy sentence. They print “Potency: 200mg/ml” in one table row and “Vial Size: 10ml” in the next. They print a figure with an asterisk and the qualification three lines lower. Writing that down as a quotation means choosing what to do with the gap, and the wrong choice — closing it — is the natural one.
9 of 18
quotations withdrawn and replaced
6 of 10
sellers whose quotations we had reconstructed
0
milligram figures that changed as a result
18 of 344
prices that carry a milligram quotation at all
All 9, with what we published and what the page printed
| Seller | What drifted | What we published, and withdrew | What the seller’s page printed |
|---|---|---|---|
| Bloom MD | Clause added | 2,400 mg total NAD+ per 12 mL bottle (200 mg/mL) | “2,400 mg total NAD+ per 12 mL bottle” |
| LaSara Medical Group | Spliced | 200mg/ml, 10ml vial | “Potency: 200mg/ml” “Vial Size: 10ml” |
| Levers Health | Footnote folded | 1200 mg of NAD+ per month (1000 mg/mo in CA, SC & IN) | “1200mg of NAD+ per month*” “*1000mg/mo in CA, SC & IN” |
| Levers Health | Footnote folded | 1200 mg of NAD+ per month (1000 mg/mo in CA, SC & IN) | “1200mg of NAD+ per month*” “*1000mg/mo in CA, SC & IN” |
| Levers Health | Footnote folded | 1200 mg of NAD+ per month (1000 mg/mo in CA, SC & IN) | “1200mg of NAD+ per month*” “*1000mg/mo in CA, SC & IN” |
| Levers Health | Footnote folded | 1200 mg of NAD+ per month (1000 mg/mo in CA, SC & IN) | “1200mg of NAD+ per month*” “*1000mg/mo in CA, SC & IN” |
| MadeMed | Re spaced | 7,500 mg per bottle (500 mg/mL × 15 mL), 30-day supply | “NAD+ Nasal Spray, 7,500mg bottle (500mg/mL × 15mL), 30-day supply” |
| Remi Meds | Re ordered | 200 mg, 30 count troches | “NAD+ Troche - 200 mg Troches 30 Count” |
| VitaStir | Re ordered | 20,000 mcg per dose, 20 injections per kit | “20 Activated NAD+ Injections (Nicotinamide Adenine Dinucleotide 20,000 mcg per dose)” |
The middle column is ours: what this site published inside quote marks and took down. It is not a quotation of any seller and is printed here without quote marks for that reason. The right column is the seller’s own text, and where it is two entries the seller printed the two parts in two places — that separation is the record, and joining them is the fault this page is about. The withdrawn strings were recovered from this repository’s own history by diffing, not from memory.
4 of the 9 are the same seller and the same sentence, because Levers Health publishes four prices that share one dose line. The asterisk in its text points at a state-by-state qualification — a lower monthly total in three states — and folding that into a parenthesis inside the quote made a qualification the seller had deliberately set apart look like part of the claim. The figure was right in every one of the four.
The 5 ways a citation drifted
Naming the ways it went wrong is more useful than counting them, because each one is a habit rather than an accident, and a habit repeats. The order below is the precedence the record uses: a citation with more than one thing wrong with it is filed under the first that applies, so the buckets always add to 9.
| What drifted | Cases | What it means |
|---|---|---|
| Spliced | 1 | two separate places on the seller’s page — a label and its value, a figure and its footnote — joined into one string, manufacturing a sentence the seller never printed |
| Footnote folded | 4 | an asterisked footnote folded into a parenthesis inside the quotation, so a qualification the seller set apart appeared to be part of the quoted phrase |
| Clause added | 1 | a clause added inside the quote marks that did not sit beside the quoted phrase on the page |
| Re ordered | 2 | the clauses re-ordered or re-worded from the order the page printed them in |
| Re spaced | 1 | whitespace inside the figure changed — “7,500 mg” where the seller prints “7,500mg” |
Counted from the record, in the declared precedence order. A kind with no records is not offered. The definitions are fixed in one place, so this page and the file agree by construction rather than by care.
What checks it now, and what that check cannot see
Every element of every milligram quotation is now checked to be a contiguous substring of a saved copy of the seller’s own page, at that offer’s own source URL, whitespace-normalised and nothing else. It reads those files raw: stripping tags out of a price page eats the very characters a quotation turns on, and that mistake has produced confident false accusations elsewhere in this network.
The quotation is stored as a list rather than a sentence, which is the structural half of the fix. Where a seller prints the two halves of a total in two places, the record keeps them in two places. There is no comma available to join them with, so the fault that produced four of these nine cannot be made again without changing the shape of the file.