How to Choose an Ashwagandha Supplement

Ashwagandha is one of the very few traditional botanicals with a real body of randomized, placebo-controlled human research behind it. That is precisely why the shelf is such a mess. Demand arrived faster than any shared standard for what a capsule has to contain, and "ashwagandha 500 mg" on a bottle is a statement about weight, not about content. Two products carrying the identical milligram number can differ enormously in the compounds the research actually measured. Everything useful on this page is about reading past that number.

Two things on the label decide most of it. The first is which part of the plant the extract came from. Traditional Ayurvedic preparation uses the root, and the human trials people cite when they talk about ashwagandha have overwhelmingly been run on root extracts. The leaves carry a different withanolide profile, and more of it by weight, which makes leaf material a cheap way to push the withanolide percentage on a label upward without matching the material that was studied. The second is the standardization itself. Withanolides are the characterized active fraction, and a percentage tells you what proportion of the powder in that capsule is that fraction. Without a percentage, the milligram figure cannot be compared to another product or to a research dose.

Everything else follows from those two. Dose only makes sense against the standardization: 600 mg of a root extract at roughly 5% withanolides and 250 mg of a root-and-leaf extract at 10% withanolide glycosides are both defensible, and neither number transfers to the other product. Format matters for the same reason — capsules and tablets carry a research-grade dose comfortably, powders can too if you buy a standardized extract rather than ground root, and gummies rarely do once you account for how little dry material fits in one. The timeline matters most of all. The trials reporting changes in perceived-stress scores and morning cortisol ran eight weeks or longer, so this is a background shift rather than something you feel on a Tuesday. Worth saying plainly while you are deciding what to spend: many of those trials were small, and several were funded by extract manufacturers. The direction is consistent across them; treat the effect size with some skepticism.

What to look for

1. The panel names the plant part, and it says root

   Look for: The words "root extract" or "Withania somnifera root" in the supplement facts panel. A product that says "root and leaf" is being honest about a legitimately different preparation. A product that says nothing at all is being honest about something else.

   Root is the traditional preparation and the material behind the bulk of the human research, so a root extract is the conservative way to buy the thing that was actually studied. Leaf material carries a different withanolide profile at a higher concentration, which is why it raises a label's standardization number cheaply.

   Red flag: "Aerial parts", "whole plant", or no plant part named anywhere. If the panel will not tell you which part of the plant you are buying, assume the cheapest input went in.

2. A withanolide percentage, with the analyte named

   Look for: "Standardized to at least 5% withanolides", typical of root-only extracts, or "at least 10% withanolide glycosides", typical of root-and-leaf. The analyte is not a detail: withanolides and withanolide glycosides are different measurements, not the same number scaled up.

   Withanolides are the characterized active fraction. Percentage multiplied by dose is how much of that fraction you actually swallow, which is the only quantity that can be compared across products or against a research dose.

   Red flag: A large milligram number with no percentage anywhere on the panel, or a percentage with no analyte stated — which lets a leaf-heavy extract borrow the credibility of a root spec.

3. The dose matches the extract it is printed on

   Look for: 300-600 mg per day for a root extract standardized around 5% withanolides, or 125-250 mg per day for the more concentrated root-and-leaf extracts standardized to withanolide glycosides. Either is a real dose. They are not interchangeable.

   Dose and standardization are one specification, not two. A smaller milligram number on a more concentrated extract can deliver a comparable amount of the active fraction, which is exactly why comparing bottles on milligrams alone tells you nothing about what you are getting.

   Red flag: 600 mg of a high-glycoside leaf-inclusive extract sold as "extra strength", when that kind of extract was studied at 125-250 mg. Or 250 mg of a 5% root extract sold as "the clinically studied dose" — 250 mg sits at the bottom of the studied range for a root extract, not at the 600 mg where most of the supportive trials sat, so the phrase is doing work the number cannot.

4. A named, standardized extract rather than a proprietary blend

   Look for: The extract identified by name with its standardization printed beside it. The two branded extracts behind most of the trial literature are the root-only one standardized to roughly 5% withanolides and dosed at 600 mg (KSM-66), and the root-and-leaf one standardized to at least 10% withanolide glycosides and dosed at 125-250 mg (Sensoril). The root-only spec carries the larger share of the published trial literature, most of it on everyday stress and sleep quality, with smaller bodies of work on strength and on hormonal outcomes. The root-and-leaf spec has been studied mainly on stress and sleep endpoints and leans more sedating, which some people prefer for evening use. Both are real specifications and this page does not pick between them.

   A named extract with a published spec is a testable claim — the material is defined and the research was run on that material. A blend that prints one combined milligram total for six botanicals hides how much ashwagandha is in there, which is generally the reason it exists.

   Red flag: "Proprietary stress blend, 750 mg" with ashwagandha listed somewhere inside it. You cannot dose what you cannot see, and blends are where underdosing hides.

5. A format that can physically carry the dose

   Look for: Capsules or tablets stating milligrams of extract per serving; powders that say "extract" and give a standardization rather than selling ground root; and for gummies, the extract milligrams per gummy plus how many gummies make a serving.

   With a standardized extract, the dose is the whole game, and format decides whether the dose fits. A gummy is mostly sugar or pectin and carries very little dry material, so gummies typically land far below the amounts used in research. Raw root powder needs several grams to approximate what a concentrated extract delivers in one capsule.

   Red flag: A gummy advertising a "full clinical dose" without printing extract milligrams, or a bulk powder priced far below every extract on the shelf — which is almost always unstandardized ground root.

6. Third-party testing you can actually look up

   Look for: A named third-party lab or certification, heavy-metal testing stated explicitly, and a batch certificate of analysis available on request with a batch number on the bottle to match it against.

   Ashwagandha is a root crop, and root crops concentrate what is in the soil, so heavy-metal testing is not a luxury on this particular ingredient. Mislabeling and adulteration are documented problems across botanicals generally, and the rare case reports of liver injury linked to ashwagandha products have included some where the product itself was in question.

   Red flag: "Lab tested" on the front of the bottle with no lab named, no COA, and no batch number. That phrase is unregulated and costs nothing to print.

What the research uses

Standardized root extract

300-600 mg/day at 5% withanolides or higher, taken with food. 600 mg is where most of the supportive trials sat.

Split option

300 mg twice daily with breakfast and dinner. Equally valid, and gentler on the stomach.

Studied range

250-600 mg/day of a standardized root extract. Higher doses have not been shown to work better.

Root-and-leaf extracts

125-250 mg/day at 10% withanolide glycosides or higher. A different specification, not a smaller version of the same dose.

Unstandardized root powder

Several grams to approximate an extract dose, with no way to verify withanolide content. Not comparable to an extract.

Time before judging it

8 weeks minimum. Sleep-quality outcomes in the research typically ran 6-12 weeks.

Cycling

8-12 weeks on, 2-4 weeks off is convention rather than evidence. Its real value is diagnostic — a break tells you whether it is still doing anything.

Claims that mean nothing

“A higher withanolide percentage means a stronger, better product.”

The percentage means nothing without the analyte and the dose beside it. Root-only extracts are typically specified around 5% withanolides; root-and-leaf extracts reach 10% or more of withanolide glycosides partly because leaf material is richer in them and partly because it is a different measurement. That is a formulation difference, not a potency league table — and the trials people are citing were run at specific doses of specific extracts, not on whichever number was largest.

“10:1 concentrated extract — ten times stronger.”

A concentration ratio describes how much raw material went into the process, not what came out of it. Without a withanolide percentage it is unverifiable, and ratios are not standardized across manufacturers or reliably consistent between batches. A stated percentage is a specification. A ratio is a story about the factory.

“Full-spectrum whole-herb ashwagandha, closer to the traditional preparation.”

"Whole herb" usually means leaf is included, which is further from the traditional preparation rather than closer — traditional use is root. "Full-spectrum" is an unregulated term with no shared definition; it can describe a genuinely well-made extract, and it can equally describe ground plant material. Only the panel tells you which one you are holding.

“Fast-acting. Calm within the hour.”

That is not what the research describes. Trials reporting meaningful changes in perceived-stress scores and morning cortisol ran eight weeks or longer, and sleep-quality outcomes generally showed up over six to twelve weeks. If a product reliably makes you drowsy on the first night, the honest question is what else is in it. For same-day support, L-theanine at 100-200 mg works through an unrelated mechanism on an unrelated timescale.

“Natural testosterone booster.”

Trials in stressed or overweight men have reported modest increases in total testosterone over eight to sixteen weeks, and the most plausible route is indirect — less cortisol load and better sleep, both of which support the system that makes it. That is a handbrake coming off, not a booster, and it is not a sensible basis for choosing one bottle over another.

“Supports adrenal fatigue recovery.”

Adrenal fatigue is not a recognized clinical entity, and a label leaning on it is telling you more about the company than about the capsule. What the trials actually measured is morning serum cortisol and validated perceived-stress scores in adults under everyday stress. That is a narrower claim, and a considerably more honest one.

The label checklist

• The supplement facts panel names the plant part, and it says root.
• A withanolide percentage is printed, and the analyte is named — withanolides, or withanolide glycosides.
• Milligrams of extract are stated per serving, and the serving size is one you will realistically take every day.
• The dose fits the extract: roughly 300-600 mg for a 5% root extract, 125-250 mg for a high-glycoside root-and-leaf extract.
• The ashwagandha is listed on its own, not buried in a proprietary blend that hides its individual amount.
• It is an extract, not ground root — or if it is a powder, the label proves the powder is a standardized extract.
• If it is a gummy: extract milligrams per gummy are printed, the full serving reaches a researched amount, and the sugar per serving is something you are willing to take daily for eight weeks.
• A third-party lab or certification is named, and heavy-metal testing is stated rather than implied.
• There is a batch number and an expiry date, and a certificate of analysis you can request for that batch.
• The claims on the bottle are structure/function — supports, helps maintain. A label claiming to treat or cure anything is telling you how the company operates.
• Nothing sedating has been quietly added. If a product promises you will feel it the first night, read the rest of the ingredient list before believing the ashwagandha did it.
• Cost per day is calculated at the dose you will actually take, not at the smallest serving printed on the label.

Before you buy

• Everything on this page assumes an adult reader, 18 and over. Ashwagandha is not something to give a child or a teenager, and questions about a young person's stress or sleep belong with their clinician.
• Do not take ashwagandha if you are pregnant. It has traditional use as an abortifacient, and it is one of the few supplements where the caution is emphatic rather than precautionary. If you are nursing or trying to conceive, talk to your clinician before you start.
• Ashwagandha can affect thyroid hormone levels. If you have a thyroid condition or take thyroid medication, this needs your prescriber's input before you start, not after.
• If you take sedatives, or any medication that already makes you drowsy, ashwagandha may add to that effect. Ask your pharmacist before combining them.
• If you have an autoimmune condition or take immunosuppressant medication, check with your clinician first — ashwagandha has immune-stimulating activity that may work against that treatment.
• It may also compound blood-pressure medication and diabetes medication. Bring the bottle to your pharmacist alongside everything else you take, rather than asking about one product at a time.
• Ashwagandha is a member of the nightshade family, alongside tomato, potato, and pepper. If nightshades reliably disagree with you, that is a plausible reason this might too.
• Case reports of liver injury have been associated with ashwagandha products. It appears rare, and some cases involved products of questionable quality, but it is real. Stop taking it and see a clinician promptly if you develop jaundice, dark urine, unusual fatigue, or upper-right abdominal pain.
• This supports everyday stress and sleep quality in adults under load. If stress or low mood is disrupting your life, that belongs with a clinician — and an herb is never a substitute for a prescribed medication.

Relevant stacks

• Stress & Mood Stack
• Sleep Stack

Frequently asked questions

What should I look for on an ashwagandha label?

Three things, in this order: the plant part, and root is what most of the research used; a withanolide percentage with the analyte named; and a dose that matches that standardization. If all three are printed, you can compare two products honestly. If any one of them is missing, you are comparing the weight of powder in a capsule and nothing more.

How many milligrams of ashwagandha should I take a day?

300-600 mg per day of a root extract standardized to at least 5% withanolides, taken with food. 600 mg is where most of the supportive trials sat, and splitting it into 300 mg twice daily is equally reasonable and slightly easier on the stomach. The concentrated root-and-leaf extracts are dosed differently, at 125-250 mg per day. Higher is not better here — the studied range tops out around 600 mg, and consistency matters far more than the number.

Is root-only ashwagandha better than root and leaf?

Better is the wrong frame. Root-only is the conservative choice because it matches both the traditional preparation and the bulk of the human trial literature. Root-and-leaf extracts are legitimate, widely used, and standardized to a higher withanolide glycoside content, which is exactly why they are dosed so much lower. The real mistake is not picking one — it is comparing their milligram numbers as though they measured the same thing.

What is the difference between KSM-66 and Sensoril?

They are the two branded extracts behind most of the trial literature, prepared differently enough that their doses differ by a factor of three or more. KSM-66 is root-only, standardized to roughly 5% withanolides, and dosed at 600 mg per day, with most of its published trials on everyday stress and sleep quality, smaller ones on strength and recovery, and early, inconsistent work on hormonal and cognitive endpoints. Sensoril uses root and leaf, is standardized to at least 10% withanolide glycosides, is dosed at 125-250 mg, and leans more sedating, which suits evening use for some people. Neither number tells you anything about the other product.

Do ashwagandha gummies work?

It depends entirely on whether the gummy carries a real dose, and most do not. A gummy is mostly sugar or pectin with a small amount of dry material in it, so gummy products commonly land well below 300 mg of standardized extract per serving, and many do not print the extract milligrams at all. If one states a standardized extract, prints the milligrams, and reaches a researched amount within a serving you will actually take, it is a perfectly good format. Check the sugar while you are reading the panel.

Capsule or powder?

Either, as long as it is a standardized extract. Capsules make the dose unambiguous, which is most of what you are paying for. Bulk powder is cheaper, but the cheap bulk powder is usually unstandardized ground root, and matching an extract dose with raw root takes several grams with no way to verify what is in them. If you buy powder, buy a standardized extract powder and weigh it rather than using a scoop.

How long before I know whether it is working?

Eight weeks. Some people notice calmer sleep and less reactivity inside the first two weeks, but the trials that reported meaningful changes in stress scores and morning cortisol ran at least eight weeks, and sleep-quality outcomes often ran longer. Judging this herb at day ten is judging it before the research says anything has happened. If you want something that works the same afternoon, buy a different compound.

Should I cycle ashwagandha?

Cycling is convention rather than evidence. The trials ran eight to twelve weeks continuously with no documented need for breaks. A common pattern is eight to twelve weeks on and two to four weeks off, and its genuine advantage is diagnostic — a break tells you whether it is still doing anything for you. If you notice a flattening of emotional range, which a minority of people do and dislike, stopping for a while is particularly informative.

Written for adults aged 18 and over. This content is for education only and is not medical advice. Supplements are not intended to diagnose, treat, cure, or prevent any disease. Talk with your clinician before starting any supplement, especially if you are pregnant, nursing, taking medication, or managing a health condition. Supplements for children are a pediatric clinician's decision — keep all supplements out of reach of children.

Our medical advisor is Dr. Garrett Wdowin, NMD, ABAAHP, FAARM, naturopathic doctor. This page has not been medically reviewed.