One claim, counted properly
Each roster is a single NSF claim search, fetched, parsed into distinct models
and checked against NSF’s own reported total before publication.
That last step is not decorative. A parser that silently returned zero rows
against a reported 145 was caught by exactly this check, and the counts on this
site are only trustworthy because it runs every time.
Counts are distinct model designations, deduplicated by company and model.
NSF lists a model once per facility and per claim, so its row totals run higher —
lead returns 1,003 rows and resolves to 615 models.
What the counts actually tell you
The spread is two orders of magnitude
Certified fields are not comparable in size, and the difference decides whether
you have a choice or a single option. Cyst reduction covers 1,104 distinct
models and microplastics 806, while nitrate has 84, microcystin 48, radon 30
from two companies and Filter First 28 from one.
A large field moves the decision to cartridge cost and form factor. A small one
means the claim itself is the constraint, which is better discovered before
comparing products than after.
Some claims exist in exactly one shape
Where a claim follows a technology rather than a material, the product form
comes with it. Every model certified for fluoride, nitrate or hexavalent
chromium is plumbed in to a separate tap, because all three are NSF/ANSI 58
claims and reverse osmosis is installed under a sink. No pitcher, faucet mount,
refrigerator filter or point-of-entry unit holds any of them, so for those three
contaminants the shape of the answer is fixed before shopping starts.
Whole-house is nearly empty of health claims
Across every roster here, point-of-entry products hold four trihalomethane
claims, the same four for VOC, and 114 softeners certified for radium and
barium. Lead, PFAS, cyst, fluoride, nitrate, hexavalent chromium, arsenic,
turbidity and microplastics return none at all. Chlorine and taste-and-odour
claims under NSF/ANSI 42 are the exception, which is why whole-house systems
are useful for aesthetics and largely unavailable for health claims.
Choose which roster to open
Open a roster if you have a contaminant and want to know whether a certified
option exists, and in what form.
Skip the rosters if you have not yet established a number. A count of
certified models answers nothing without something to treat.
Stop entirely if the roster shows no certified product in a form you can
fit. That is a complete answer, and it is cheaper than the alternative.
The rosters
| Claim |
Certified models |
Notable |
| Lead |
615 |
Six product forms; 282 are fridge filters |
| Cyst |
1,104 |
The largest field on the site |
| PFAS |
72 |
PFOA and PFOS only; no GenX claim exists |
| Trihalomethanes |
153 |
Includes four point-of-entry models |
| VOC |
146 |
One claim covering 53 named contaminants |
| Nitrate |
84 |
One product form only |
| Fluoride |
127 |
One product form only |
| Arsenic |
135 |
Every claim is pentavalent-only |
| Chloramine |
137 |
Separate claim from chlorine |
| Whole house |
Few |
Almost no health claims certified |
| Shower filters |
6 |
One claim, aesthetic effects only |
Bottled water is the other comparison
Not a certified product in this sense, and regulated by FDA rather than EPA.
→ Bottled water versus tap: who regulates what
Before opening any of them
→ What is in city water: the regulated, the advisory, and the unmeasured
→ How to check a certification claim yourself
→ Which claims exist in which product form
→ The regulated contaminants with no certified filter
The bottom line
Use a roster to answer one question: does a certified option exist for my contaminant, and in a form I can fit? The counts vary by two orders of magnitude and several claims exist in one product form only. Where a field is small, that is worth knowing before shopping rather than after. Every figure here is re-checked against NSF's listings by a drift script, so a count that changes is caught rather than quietly aging.
Who this is not for
These are counts of NSF listings on a stated capture date, not recommendations and not a market survey. Counts are distinct models deduplicated by company and model, so they sit below NSF's reported row totals. WQA and IAPMO certify separately under the same standards and are not counted here. No product on this site is endorsed, and no affiliate relationship exists.
Questions
What is a roster on this site?
One NSF claim search, fetched and parsed into distinct models. The extracted row count is checked against NSF's own reported total before anything is published, so a parsing failure shows up as a mismatch rather than as a wrong number.
Why are the counts lower than NSF's?
Because NSF lists a model once per facility and per claim, while these counts are distinct model designations deduplicated by company and model. Lead returns 1,003 rows, for example, resolving to 615 distinct models.
Which contaminant has the most certified filters?
Of the claims rostered here, cyst reduction is the largest field at 1,104 distinct models, followed by microplastics at 806 and lead at 615. Nitrate, by contrast, has 84 and every one is the same product form.
Which fields are surprisingly small?
Radon has 30 models from two companies, Filter First 28 from one, and shower filters 6 under a standard offering a single claim. Small fields are worth knowing about before shopping, not after.
Are the counts kept current?
A drift script re-runs every recorded claim query against NSF's listings and compares the result with the stored count. It currently covers 45 queries. A changed count is flagged rather than left to age quietly on a page.
Do these pages recommend products?
No. They count what is certified and report the structure of each field. No affiliate programme is approved, so no paid link exists anywhere on this site, and no model is endorsed.
Why do some claims exist in only one product form?
Because the claim follows the technology. Fluoride, nitrate and hexavalent chromium are NSF/ANSI 58 claims, and reverse osmosis is installed under a sink with its own tap, so no other form holds them.
What about WQA or IAPMO certified products?
They certify to the same ANSI standards and publish their own listings, which are not counted here. A product absent from these rosters may be certified elsewhere, and every roster says so.
Sources — 12 claims
- NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 53 lead reduction claim search (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF’s listing search returned 1,003 rows for lead reduction under NSF/ANSI 53 on 2026-09-10, resolving to 615 distinct model designations from 63 companies. (captured 2026-09-10)
Limits: NSF lists some models on several rows, so the row count and the distinct-model count measure different things and both are reported. NSF listings only; WQA and IAPMO certify to the same standards and publish separately. - NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 58 nitrate/nitrite reduction claim search (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF listings returned 84 rows for nitrate/nitrite reduction under NSF/ANSI 58 on 2026-09-10, resolving to 84 distinct model designations from 28 companies, every one listed as plumbed-in to separate tap. (captured 2026-09-10)
Limits: One category in one certifier listing on one date. EPA maps nitrate to distillation, reverse osmosis or ion exchange, and only the reverse osmosis standard offers a searchable nitrate claim, so the single category here reflects how NSF indexes claims as much as what treats nitrate. - NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 53 PFOA, PFOS and Total PFAS reduction-claim searches (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
PFOA, PFOS and total PFAS reduction are selectable in NSF’s listing search under NSF/ANSI 53 only, and no PFAS claim appears under NSF/ANSI 58, although EPA advises looking for certification under NSF/ANSI 53 or NSF/ANSI 58 for PFAS reduction. (captured 2026-09-10)
Limits: A statement about which claims NSF offers, not about what reverse osmosis does. A system certified to NSF/ANSI 58 may also hold NSF/ANSI 53 certification separately, which is where any PFAS claim it carries would sit. The tension with EPA’s advice is reported, not resolved. - NSF Official Listings — Drinking Water Treatment Units, product category headings across fourteen claim searches (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
Certified claims are distributed very unevenly across product forms on 2026-09-10. Of 615 models certified for lead reduction under NSF/ANSI 53, 282 are refrigerator filters, 161 plumbed-in, 113 plumbed-in to a separate tap, 29 faucet mount, 24 counter-top and 4 pour-through pitchers. The refrigerator filter carries 45.9 percent of the entire lead-certified field; the pitcher carries 0.7 percent. (captured 2026-09-11)
Limits: Counts are distinct model designations, deduplicated by company and model. NSF lists a model once per facility and claim, so NSF's own reported row counts are higher. The category is NSF's printed table heading, not an inference about installation. Share of a certified field is not a statement about performance, price or availability, and a form with few listings may still contain the right product. - NSF Official Listings — Drinking Water Treatment Units, product category headings across fourteen claim searches (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
Every model certified under NSF/ANSI 58 for fluoride reduction (127), nitrate and nitrite reduction (84) and hexavalent chromium reduction (120) is listed in a single product form on 2026-09-10: plumbed-in to a separate tap. No pitcher, faucet mount, counter-top, refrigerator filter or point-of-entry product holds any of the three. (captured 2026-09-11)
Limits: Counts are distinct model designations, deduplicated by company and model. NSF lists a model once per facility and claim, so NSF's own reported row counts are higher. The category is NSF's printed table heading, not an inference about installation. These are reverse osmosis claims and reverse osmosis is installed under a sink with its own faucet, so the form follows the technology. It bounds where a certified product exists, not where the contaminant can be treated. - NSF Official Listings — Drinking Water Treatment Units, product category headings across fourteen claim searches (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
Point-of-entry products are almost absent from the health-claim fields on 2026-09-10. Across fourteen claim rosters the only point-of-entry listings are 4 models for trihalomethanes, the same 4 for VOC, and 114 softeners certified for radium 226/228 and barium under NSF/ANSI 44. Lead, PFAS, cyst, fluoride, nitrate, hexavalent chromium, arsenic, turbidity and microplastics return no point-of-entry product at all. (captured 2026-09-11)
Limits: Counts are distinct model designations, deduplicated by company and model. NSF lists a model once per facility and claim, so NSF's own reported row counts are higher. The category is NSF's printed table heading, not an inference about installation. Bounded to the fourteen claims these rosters cover. NSF/ANSI 42 chlorine and taste-and-odour claims do list point-of-entry products, and those are counted in the separate point-of-entry census rather than here. - NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 53 microcystin reduction claim search (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF listings returned 52 rows for microcystin reduction under NSF/ANSI 53 on 2026-09-10, resolving to 48 distinct models from 7 companies: 31 plumbed in to a separate tap, 16 counter-top and 1 refrigerator filter. No pitcher, faucet mount or point-of-entry product holds the claim. (captured 2026-09-11)
Limits: Microcystin is the only cyanotoxin with an NSF reduction claim. Cylindrospermopsin and anatoxin-a have none, and microcystins are a family of variants while the claim appears under a single name. - NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 53 radon reduction claim search (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF listings returned 34 rows for radon reduction under NSF/ANSI 53 on 2026-09-10, resolving to 30 distinct models from only 2 companies — 15 counter-top and 15 plumbed in to a separate tap. Every one of the 30 models holds an identical set of 11 claims, so the radon claim exists on a single product family rather than across a market. (captured 2026-09-11)
Limits: A certified claim for a contaminant with no enforceable limit. Certification records tested performance, and says nothing about whether treating water is the useful response when EPA puts only 1 to 2 percent of indoor radon exposure down to drinking water. - NSF Official Listings — Drinking Water Treatment Units, asbestos reduction claim searches (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF listings returned 364 rows for asbestos reduction under NSF/ANSI 53 on 2026-09-10, resolving to 264 distinct models from 32 companies, and 10 rows under NSF/ANSI 58 resolving to 10 models from 2 companies. Not one of the 264 NSF/ANSI 53 models is a pour-through pitcher. (captured 2026-09-11)
Limits: EPA regulates fibres longer than 10 micrometers; NSF listings state a claim name and not a fibre length, so the listing does not record which sizes a model was tested against. A pitcher without the claim is untested against asbestos, not shown to fail. - NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 42 Filter First claim search (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
The Filter First claim sits under NSF/ANSI 42, the standard NSF describes as covering non-health-related contaminants, and no lead claim appears anywhere in the Filter First listing. Every one of the 28 models holding it does separately hold a lead reduction claim under NSF/ANSI 53, so the lead performance those products carry comes from the NSF/ANSI 53 claim and not from the Filter First claim. (captured 2026-09-11)
Limits: A statement about which claim carries which performance, not a criticism of the products. All 28 hold both. The point is that a purchaser checking only for the Filter First name would not thereby have checked for a lead claim, because the two are recorded separately and under different standards. - NSF Official Listings — Drinking Water Treatment Units, listing table column headers (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF listings record a replacement element, a rated service cycle in gallons and a flow rate in gallons per minute for each filter model, and for reverse osmosis systems a replacement module and a daily production rate in gallons per day. (captured 2026-09-10)
Limits: The rated service cycle is the volume the claim was validated across, recorded per model. NSF separately states capacity and rated service cycle claims are not applicable for mechanical filtration, so sediment filters carry no such figure. - NSF Official Listings — Drinking Water Treatment Units (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF Official Listings separate systems from components, noting under NSF/ANSI 42 that all replacement elements are components and under NSF/ANSI 58 that all replacement modules are components, and listing component categories such as filter cartridges and filter housings separately from systems. (captured 2026-09-10)
Limits: A component listing describes the part, not the assembled system it goes into. Whether the finished system is certified is a separate listing to look up.