The efficacy evidence for stress and sleep is actually reasonable for a botanical. What the marketing hides is a real, if uncommon, signal of ashwagandha-associated liver injury serious enough that a national regulator has issued an alert. Use it deliberately, short-term, and stop if anything feels off.
The one-paragraph version
Multiple meta-analyses find ashwagandha reduces perceived stress and anxiety and improves several sleep measures versus placebo, with the clearest effects at ~300–600 mg/day of standardized root extract for 8+ weeks S1S2S3. That's a genuinely better evidence base than most "calm" supplements have. But since 2017 there have been dozens of case reports of ashwagandha-associated liver injury, most mild-to-moderate and reversible but a few severe, and Australia's regulator flagged it in 2024 S4S5. This is not the consequence-free herbal the influencer market implies. A short, monitored trial is reasonable for healthy adults; long-term daily use and any use with liver risk factors is where caution turns into a hard stop.
What it is and how it works
Ashwagandha (Withania somnifera) is an adaptogenic botanical from Ayurvedic tradition. Its activity is attributed largely to withanolides, steroidal lactones; standardized extracts are dosed by withanolide content S1. Proposed mechanisms center on modulating the stress response, including lowering cortisol S1.
What the evidence actually supports
Stress and anxiety — decent for a botanical. A 2024 systematic review and meta-analysis found ashwagandha significantly reduced perceived stress and anxiety versus placebo, including a meaningful drop on the Perceived Stress Scale S2. NIH's fact sheet agrees the human evidence is suggestive of benefit while noting trials are short and varied in quality S1.
Sleep — modest, better in insomnia. A meta-analysis of sleep trials found benefits across sleep quality, onset latency, total sleep time, and efficiency, with more prominent effects in people diagnosed with insomnia, at doses ≥600 mg/day, and durations ≥8 weeks S3. A reasonable, evidence-backed option for stress-related poor sleep.
Testosterone and body composition — weaker, thinner. Small trials report increases in testosterone and training adaptations, but these rest on limited, often industry-linked studies and shouldn't be treated as established S1. Promising-but-unproven, not a reason to buy.
Who actually benefits
Healthy, stressed adults looking for a short-term aid for perceived stress or stress-related sleep are the best-supported use S1S2S3. It is not appropriate for people with liver disease or liver risk factors, pregnancy/breastfeeding, thyroid disorders (it can affect thyroid hormones), or those on immunosuppressants or hepatotoxic drugs S1S4.
Dosing (standard, well-established)
Trials most commonly use 300–600 mg/day of a standardized root extract, often split into two doses, for 8–12 weeks S1S2S3. Longer-term daily use is not well studied — a time-limited trial with breaks is the more defensible pattern given the liver-safety uncertainty. Standardization to withanolide content matters more than raw milligrams.
Safety
Short-term use in trials is generally well tolerated, with mild GI upset and drowsiness the usual complaints S1. The real concern is the liver: LiverTox classifies ashwagandha as a cause of clinically apparent liver injury, and case series and reports — dozens since 2017, with Australia's TGA issuing a 2024 safety alert after multiple reports including hospitalizations — describe injury that is usually reversible on stopping but occasionally severe S4S5. The true incidence is unknown because the denominator (how many users) isn't. Practical rules: don't stack it with alcohol or other hepatotoxic supplements, avoid it entirely with existing liver disease, and stop and seek care for jaundice, dark urine, right-upper-abdominal pain, or unusual fatigue.
The marketing myths
- "Natural, so it's safe." The clearest myth here — a documented liver-injury signal says otherwise S4S5.
- "Take it every day indefinitely for calm." Long-term safety isn't established; short, monitored courses are the defensible pattern S1.
- "It'll raise your testosterone and build muscle." The evidence is small and often industry-linked — not established S1.
- "All ashwagandha is equivalent." Effects track standardized withanolide content and extract type, not just the milligram number on the front.
Sources
Every reference below is a primary source cited in this guide.