This is a site about mold and indoor air quality, and water is, on its face, a different topic. We’re including it anyway, because the broader category we actually care about — what’s in the environment inside your house that’s affecting the people in it — doesn’t stop at the spore count and the VOC reading. Drinking water is one of the larger, more neglected pieces of that picture, and since the Flint crisis dragged the subject back into mainstream awareness, the gap between what homeowners worry about and what they actually do about it has only widened.

The pattern we see is bimodal. Some households ignore tap water entirely — drink it, cook with it, never think about it. Others spend $4,000 on a whole-house reverse-osmosis system they didn’t need to address a problem they didn’t actually have, sold to them by a salesperson in a polo shirt with a TDS meter. Both of these are the wrong answer. The right answer is dull and specific: figure out what’s actually in your water, match the filter to the actual contaminants, and stop there.

This article is the calibrated version. No affiliate links to $2,000 systems, no fear-marketing about “toxins,” and no pretending the answer is the same for everyone. Your water source determines most of it.

Start by knowing your water source

There are two regimes, and almost everything that follows depends on which one you’re in.

Municipal / utility water

If a water utility serves your house, it is required by the Safe Drinking Water Act to test for ~90 regulated contaminants and to publish an annual Consumer Confidence Report (CCR) — sometimes called a Water Quality Report — listing what was detected and at what levels relative to the EPA’s maximum contaminant levels (MCLs).

Find yours. It’s usually mailed once a year and posted on the utility’s website. The EPA also maintains a drinking water overview that explains how the system works and links to state-level resources. For a more accessible read of the same data, the EWG tap-water database aggregates CCRs by ZIP code and flags contaminants that exceed EWG’s health guidelines, which are stricter than the EPA’s by design (more on that below).

The CCR is step zero. Most of the decisions in this article — do I need a filter at all, which one, certified to what — flow directly from what your CCR shows. Reading the report for ten minutes is worth more than any amount of generic filter shopping.

Private well water

If your water comes from a private well, no agency is testing it for you and no report is mailed to you. You are the utility. The EPA and CDC recommendation, roughly:

  • Annually: test for total coliform bacteria, E. coli, nitrates, and pH. These are the fast-moving indicators of a problem with the well itself.
  • Every few years, or after any change in water quality: a broader panel for metals (arsenic, lead, manganese), VOCs, and pesticides — weighted toward whatever your local geology and land use suggest. State extension offices typically publish a recommended panel for your county.
  • After any flood, well work, or nearby construction: retest the basics.

Well water gives you no baseline guarantees and a lot more responsibility. The upside is that, once you know what’s in it, the fix is often a targeted appliance — a UV sterilizer for microbial issues, a specific media for arsenic — rather than a sweeping system.

The contaminants worth caring about (and why)

Not every detection on a CCR is meaningful. Here are the ones that actually move the needle, in rough order of how often we see them mishandled.

Lead

There is no safe level of lead exposure. EPA’s action level is 15 ppb at the tap, but the health-based goal is zero, and child neurodevelopment is affected at much lower exposures. Lead in tap water almost never comes from the source water or the utility’s main; it comes from your service line, your solder, or your fixtures. Houses built before 1986 (when lead solder was banned in plumbing) and houses with unknown or known-lead service lines are the at-risk population. Brass fixtures — including many “lead- free” labeled ones — can leach small amounts.

If any of those apply to you, test specifically for lead. It’s cheap (~$30 for a single-sample lead test through a state-certified lab) and it’s binary: either you have a measurable problem or you don’t. We cover the response below.

PFAS (“forever chemicals”)

Per- and polyfluoroalkyl substances are the emerging contaminant of the decade. They don’t break down meaningfully in the environment or the body, and they accumulate. EPA finalized national MCLs in 2024 — 4 parts per trillion for PFOA and PFOS individually, with limits on several others. That’s an extraordinarily low number, and many utilities are still working out compliance.

If you live downstream of an industrial site, a military base with firefighting-foam history, or a manufacturing region, your water system may be affected. Your CCR and the EWG database will both flag it. Activated carbon removes a substantial portion of PFAS; reverse osmosis removes more; ion-exchange resins designed for PFAS remove the most. NSF/ANSI 53 and 58 cover PFAS in many cartridges now; 473 is the PFAS-specific certification.

Disinfection byproducts (THMs, HAAs)

When chlorine reacts with organic matter in the source water, it forms trihalomethanes (THMs) and haloacetic acids (HAAs). These are essentially universal in chlorinated systems and are associated with a modest long-term cancer risk at chronic exposure. EPA regulates total THMs at 80 ppb; EWG’s health-based guideline is much lower.

Activated carbon — any decent pitcher, faucet-mount, or under-sink filter — handles most of these. This is the most common reason basic filtration makes sense in a typical municipal household: not because the water is dangerous, but because cheap filtration drops a chronic low-level exposure meaningfully.

Nitrates

Largely an agricultural-runoff problem and a well-water problem, though it appears in some municipal systems near farms. EPA’s MCL is 10 ppm. The acute risk is to infants — “blue baby syndrome” (methemoglobinemia). If you have a well in a farming area and infants in the house, test for nitrates and take any positive result seriously. Activated carbon does not remove nitrates; reverse osmosis or ion exchange does.

Arsenic

Regional, mostly geological, mostly well water. Parts of New England, the Upper Midwest, the Southwest, and pockets elsewhere have natural arsenic in groundwater. EPA’s MCL is 10 ppb; chronic exposure above that is associated with cancer and cardiovascular disease. If your area has any arsenic history at all, test for it. Reverse osmosis works; specialty iron-oxide media works better for high-concentration well water.

Microbial contamination

E. coli and total coliforms in a well sample mean surface water or sewage is reaching the well — a structural problem that needs fixing at the source (well cap, septic distance, casing integrity) before any filtration matters. Legionella is a building-plumbing issue — stagnant hot water in tanks set below 120 °F is the classic risk. UV sterilization handles microbial loads downstream of the source fix; it does nothing for chemistry.

Hardness (calcium, magnesium)

Cosmetic and operational, not a health issue. Hard water leaves scale on fixtures, reduces soap effectiveness, and shortens water-heater life. The marketing claim that “soft water is healthier” is essentially fiction. Soften if you want to protect appliances or hate scale, not for your body. And note that softened water has higher sodium — people on sodium-restricted diets sometimes drink unsoftened water at a separate tap.

Chloramine

Some utilities use chloramine (chlorine + ammonia) instead of free chlorine because it persists longer in distribution. It’s harder to remove than chlorine — basic carbon will reduce it, but catalytic carbon does it properly. Chloramine matters if you brew kombucha, keep fish, or have a sensitivity to the taste. Check your CCR for which disinfectant your utility uses.

The testing question

For municipal water, your CCR is often enough. If it shows nothing of concern at your address and you have no reason to suspect in-home plumbing issues, you may not need to test at all.

When testing makes sense:

  • You have a well. Required, see above.
  • Pre-1986 house, or unknown service line, or any lead concern. Test specifically for lead at the kitchen tap, first-draw after overnight stagnation.
  • You’ve moved into a new place and don’t trust the seller’s disclosures. A basic mail-in panel is a reasonable peace-of- mind move.
  • Your CCR shows a borderline detection and you want to know what’s actually coming out of your tap (vs. the system average).
  • Anyone in the household is pregnant, an infant, or immunocompromised, and the answer matters more.

Testing options, roughly:

  • Utility-provided free basic tests — some municipal utilities will run a free lead or basic-mineral test on request. Call and ask. Reasonable starting point.
  • Mail-in lab tests — $150–$500 depending on panel breadth. Tap Score by SimpleLab is reputable and the reports are written for humans; state extension labs and the major commercial labs (Eurofins, Pace, NTL) are also fine. For lead alone, $30. For a full “essential” panel of metals, inorganics, and disinfection byproducts, $200ish. For VOCs, PFAS, pesticides — add on as relevant.
  • In-home test strips — chlorine, hardness, pH, sometimes iron. Cheap, indicative not definitive. Useful for monitoring after you’ve established a baseline.

When the results come back, the question is not “is anything detected” — almost everything is detected at trace levels. The question is what’s above the MCL or above the more conservative health-based guideline (EWG’s, or your state’s where stricter). Anything well below either, ignore. Anything above either, that’s your shopping list for filtration.

The filter categories — match the tool to the problem

There is no single “best filter.” There are filter types, each good at certain things and useless at others. The honest exercise is to look at your contaminant list and pick the type that handles it — not the one with the most buttons.

Activated carbon

The workhorse. Comes as pitchers (Brita-class), faucet-mounts, refrigerator filters, countertop units, and under-sink cartridges. Good at: chlorine, chloramine (with catalytic carbon), taste and odor, many VOCs, disinfection byproducts, some pesticides. Some carbon-block cartridges with appropriate certifications also handle lead, cysts, and PFAS — but only if the specific cartridge is certified for it.

This is the right answer for most municipal-water households whose only real concern is chlorine taste and the chronic low-level THM exposure. Get NSF/ANSI 42 (aesthetic) and 53 (health-related) certified, change cartridges on schedule, done.

Reverse osmosis (RO)

Under-sink unit with a separate small faucet at the kitchen sink, usually with a holding tank in the cabinet. Pushes water through a semi-permeable membrane. Removes nearly everything: nitrates, fluoride, most PFAS, arsenic, lead, dissolved solids, most pharmaceuticals.

The honest trade-offs: it wastes water (typically 2–4 gallons to make 1 gallon of product, though newer “zero-waste” designs recirculate), it’s slower than tap pressure, it removes minerals that contribute to taste (some people find RO water tastes flat — remineralization cartridges exist), and the membrane is the expensive part to replace. It’s the right answer when you have a specific serious contaminant the membrane handles. It’s overkill when chlorine taste is your actual problem.

Ion exchange

The category that includes whole-house water softeners (sodium exchange for calcium and magnesium) plus specialty media for nitrate removal, perchlorate, and some PFAS. Useful and targeted; not a general-purpose filter.

UV sterilization

A tube with a UV lamp; water passes through and microbes are inactivated. Removes nothing chemical. Right answer for well-water microbial concerns after you’ve addressed the source.

Distillation

Boils water and condenses the steam, leaving virtually everything behind. Removes nearly all contaminants including ones nothing else catches. Slow, electrically expensive, removes minerals. A laboratory-grade answer; rarely the right one for a household.

Whole-house carbon

Big tank in the basement; all incoming water passes through. Right answer in a small set of cases: shower-water chlorine smell is genuinely bad, you have VOC-contaminated source water, or you have specific reasons to want carbon at every fixture. Overkill for most households whose only need is filtered drinking water at the kitchen tap.

Specialty cartridges

For specific contaminants where the standard categories don’t do the job efficiently:

  • Fluoride — activated alumina cartridges (RO also works).
  • Arsenic — iron-oxide media for high-concentration well water.
  • PFAS — dedicated ion-exchange resin in addition to carbon.

The certification question

A filter is only as good as the standards it’s certified against, not what the box claims. The relevant NSF/ANSI standards:

  • 42 — Aesthetic effects (chlorine, taste, odor, particulates).
  • 53 — Health-related contaminants (lead, VOCs, cysts, some PFAS, depending on cartridge).
  • 58 — RO systems specifically.
  • 401 — Emerging contaminants (pharmaceuticals, certain hormone-disruptors).
  • 473 — PFAS specifically.
  • P231 — Microbiological purifiers (for camping, also relevant for well-water UV pairings).

A filter labeled “NSF certified” without specifying what for is marketing. A cartridge certified to NSF/ANSI 42 only is not certified to remove lead, no matter what the marketing copy implies. The vendor’s product page often outruns the actual certifications by a wide margin.

Verify on NSF’s website directly — they publish a searchable database of certified products by standard and by contaminant. If a product you’re considering isn’t listed there for the contaminant you care about, the claim is unsupported.

The over-buying pattern

The water filter industry has become very good at fear marketing. The common patterns we see, in roughly descending order of financial damage:

  • Whole-house reverse osmosis systems ($3,000–$10,000 installed) sold to people whose only need was a $40 under-sink carbon filter for chlorine taste. RO at the point of use, where you actually drink the water, is almost always the better answer if RO is justified at all.
  • Five-stage countertop systems sold for problems a basic pitcher would have handled.
  • Alkaline ionizers ($1,500–$3,000) marketed as “healthier” on the basis of pH alone. The evidence base for health benefits from alkaline water is essentially nonexistent at the levels these machines produce.
  • TDS-meter sales theater — a salesperson dips a meter in your tap water, points to the number, and implies a high reading means contamination. TDS measures all dissolved solids including harmless minerals; it’s a near-useless proxy for health-relevant contamination.
  • “Free water test” at the door — same playbook as the free mold inspection. It’s a sales call, not a test.

The honest filter recommendation depends entirely on your CCR (or well panel) plus your specific concerns. For most municipal-water households in the U.S., the right answer is a $30–$80 pitcher or faucet-mount carbon filter with NSF 42+53 certifications, plus cartridge replacements on schedule. Anything beyond that should be justified by a specific contaminant.

Maintenance — the un-fun part

A filter you don’t change is worse than no filter at all. Spent cartridges channel — water finds the path of least resistance through the exhausted media and emerges essentially untreated. And warm, wet, organic-rich filter media is a comfortable place for bacteria to multiply; an old carbon cartridge can put out water with a higher microbial count than the tap water going in.

Set reminders. The cartridge spec sheet tells you the design life, usually in gallons or months — whichever comes first. A few rough schedules:

  • Pitcher filters — every ~40 gallons, typically 2 months for a small household.
  • Faucet-mount and refrigerator filters — every 3–6 months.
  • Under-sink carbon — every 6–12 months depending on use.
  • RO systems — pre- and post-filters quarterly to annually; membrane every 2–5 years; sanitize the holding tank when changing the membrane.
  • Whole-house carbon tanks — backwash on schedule, media replaced every several years.
  • UV lamps — replaced annually whether or not they still glow; output degrades long before they go dark.

For households with Legionella risk concerns — and these are rare in residential, but possible if there’s an immunocompromised person and a hot-water tank that runs cool — set the tank temperature to at least 120 °F (140 °F is more conservative but a scald risk), and consider a recirculation pump or periodic flush of seldom-used hot-water lines.

Lead specifically — the highest-stakes case

Lead deserves its own short section because the stakes are higher than anything else on this page and the response is well-defined.

Test if: your house pre-dates 1986, your service line is unknown or known to be lead, you have brass fixtures, or any documented lead in the area. A single first-draw test at the kitchen tap after overnight stagnation costs about $30.

While you’re waiting for results, and as ongoing habit if any risk factor applies:

  • Run cold water 30 seconds after any stagnation period (first thing in the morning, after coming home, after a long meeting) before drinking or cooking. Flushing clears the water that’s been sitting in your fixtures and service line.
  • Use cold for drinking and cooking, never hot. Hot water leaches more lead from solder and brass; the hot tap is not a shortcut for boiling.
  • Use an NSF/ANSI 53-certified lead-reduction filter if the test comes back positive at any meaningful level. Verify the specific cartridge is certified for lead, not just the brand.

If the test comes back elevated, the longer-term answer is service-line replacement — your utility may have a program that subsidizes or fully funds this; many do now under federal funding. If there are children in the house, talk to the pediatrician about blood-lead testing. Lead is the one contaminant where the precautionary response is uncontroversially worth the cost.

Bottled water — the calm take

Bottled water in the U.S. is regulated by the FDA, not the EPA, and is in many cases less rigorously tested than municipal tap water. A large fraction of it is municipal tap water that’s been filtered and bottled. It is not, in general, “purer” than what comes out of your kitchen sink through a decent filter.

It is convenient. The environmental cost — plastic, transport, disposal — is real. For a municipal-water household, a good filter is cheaper per gallon, lower-impact, and at least as safe. For a household on a well with confirmed serious contamination, bottled water is a reasonable bridge while you address the source. For travel or emergencies, sure. For daily use as a “safer” alternative to filtered tap, it’s marketing.

The water-and-mold overlap

This is the mold-and-IAQ site, so we have to close the loop. Standing water from any plumbing source — including the small chronic drips at filter housings — is mold food. The classic pattern: an under-sink RO system with a slow leak at one of the fittings, dripping into the cabinet for months, found only when the homeowner pulls out the cleaning supplies and discovers a black-stained cabinet floor.

When you install any new under-sink water treatment, and on every weekly mold check, look under every sink with a flashlight. Run a hand along the connections; feel for damp. A $5 puck-style water alarm under the sink catches it for you. The lifetime cost of treating water that then leaks and rots out your kitchen base cabinets is much higher than the cost of preventing the leak.

What to do today

A short, ordered checklist.

  1. Find your CCR (if on municipal water). Read it. Cross- reference your address on EWG’s tap-water database for context.
  2. If you have a well, schedule a basic annual panel (bacteria, nitrates, pH) plus a broader scan if you’ve never done one or it’s been several years.
  3. If your house is pre-1986, or any lead risk factor applies, order a $30 lead test and run it first-draw at the kitchen tap.
  4. Match filter type to actual contaminants. For most municipal-water households: a pitcher or faucet-mount carbon filter NSF-certified to 42+53 is the entire answer.
  5. Verify certifications on NSF’s website, not the product box.
  6. Set cartridge-replacement reminders before you install anything. The filter you forget to maintain is worse than no filter.
  7. Look under every sink for drips from any water-treatment gear, weekly. Add it to your mold-check loop.
  8. Don’t buy the whole-house RO from the door-to-door salesperson with the TDS meter. Almost no one needs it.

Boring? Yes. That’s the point. Most of what’s worth doing here costs under $100 and a Saturday morning. The expensive answers are usually the wrong answers.