There is no single “right” dose for a nootropic. The useful amount depends on the compound’s half-life, your individual sensitivity, what you stack it with, and what you are trying to achieve.
This guide covers the principles that apply across the board, then gives a reference table for every nootropic compound and nootropic herb with a monograph here. Each figure is the dose from that compound’s own page, where it sits alongside the trials it came from.
This is general educational information, not medical or dosing advice. Start at the low end, change one variable at a time, and speak to a qualified practitioner before combining nootropics with medication.
Start low, then titrate
The single most useful habit is to begin at the lowest commonly used dose, give it a few days, and only increase if you need to.
Most of the unwanted effects of nootropics — jitteriness, headaches, irritability, flattened mood — come from taking too much, too soon, or stacking several stimulating compounds at once.
A bigger dose is not a better dose. Many of these compounds follow an inverted-U, where the benefit sits in the middle of the range and more actually works worse. Sunifiram showed exactly that curve in the animal work that is all it has.
Potency also varies by orders of magnitude across compounds that get sold side by side.
A dose that is trivially small for one is a large overdose of another.
Why half-life drives your timing
A compound’s half-life is how long your body takes to clear half of a dose. It tells you two things at once: how long you will feel an effect, and how soon a second dose makes sense.
Short
1–3 hEffects arrive and fade quickly. Good for a single focused block; needs redosing for all-day coverage.
Medium
4–8 hOne morning dose covers most of a working day.
Long
10 h+One dose lasts well into the evening, so take it early. Modafinil is the clearest example.
Two Rules of Thumb
It takes roughly 4–5 half-lives to clear a compound almost completely. Caffeine, at a ~5-hour half-life, is nearly gone after a day — but a quarter of an afternoon dose is still circulating that night, which is why a late coffee costs you sleep.
Dosing more often than the half-life causes a compound to accumulate. That is intended for some, and it is how a steady level builds. For stimulants it is also how tolerance and crashes build.
A short plasma half-life does not always mean a short effect.
Noopept clears from plasma within about 25 minutes, and the activity is carried by a longer-lived metabolite. Aniracetam is the same story. Phosphatidylserine is cleared from blood in hours but works by being built into cell membranes, and its trial benefits took 6–12 weeks to appear. Creatine has no meaningful half-life at all — muscle and brain stores saturate over weeks and then stay saturated.
Where a compound works through a metabolite or a stored pool, the half-life tells you almost nothing about when to take it. Where it works directly, it tells you almost everything.
Worked example: caffeine
Caffeine has a half-life of about 5 hours — longer if you are a slow metabolizer, on oral contraceptives, or pregnant, and shorter if you smoke.
Take 100 mg at 8 a.m.:
1 p.m.
(one half-life) about 50 mg remains. That afternoon dip is partly the level falling.
+100 mg
A second 100 mg dose at 1 p.m. stacks on the leftover 50 mg, putting ~150 mg on board — a higher peak than the morning.
11 p.m.
That afternoon dose still leaves ~25 mg circulating, enough to delay sleep in sensitive people.
A smaller second dose, say 50 mg early in the afternoon, usually sustains focus better than a second full dose. A hard “no caffeine after about 2 p.m.” cutoff protects sleep.
Pairing caffeine with L-theanine at roughly 2 parts theanine to 1 part caffeine smooths the peak and takes off the edge.
What food and timing change
For several of these compounds, when and how you take a dose matters more than the number.
Take with food
Absorption depends on itVinpocetine is the extreme case: oral absorption is around 7% fasting and 60–80% with a meal, so a fasted dose is close to a wasted one. Aniracetam, pramiracetam, phosphatidylserine and sulbutiamine are all fat-soluble and absorb better with a meal containing some fat.
Take away from food
Protein competesL-tyrosine competes with other large neutral amino acids for the same transporter into the brain, so a high-protein meal blunts it. Take it 1–2 hours before the task it is meant to cover.
Take early
Because of sleepModafinil is a once-daily morning drug for exactly this reason. Caffeine, phenylpiracetam and centrophenoxine all disrupt sleep taken late in the day.
Timing is not critical
Whenever you'll rememberCiticoline, creatine, piracetam and huperzine-A can be taken with or without food, at whatever time you will actually remember.
Splitting a dose is the usual answer to a short half-life rather than a preference. Oxiracetam, piracetam, pramiracetam, vinpocetine and acetyl-L-carnitine are all split across the day in the studies that used them, and for ALCAR there is a second reason: absorption is saturable, so a larger single dose delivers proportionally less.
When a compound has no established dose
Twelve of the compounds below have no dose to look up, and the tables say so rather than filling the gap. The reasons differ, and they are worth telling apart.
Nothing has ever been measured in a person.
Dihexa, P21, N-Acetyl Selank Amidate and N-Acetyl Semax Amidate have never been administered to a human in a published study. There is no bioavailability figure and no half-life from which a dose could be built.
Only animals have been dosed.
A dose exists, but it is not a dose that worked.
Fasoracetam was titrated to 400 mg twice daily in ADHD trials that missed their endpoints. Coluracetam has one figure on record, from a discontinued depression trial with no published outcome. These are protocol doses, not effective doses. DMAE is the same problem in supplement form: the 100–300 mg on the label is marketing convention, and the controlled cognitive trials that did test deanol were negative.
The dose belongs to another route or another molecule.
For a buyer
The practical consequence is the same in every case. The number on the label was chosen by the seller. There is no trial behind it, no maximum tolerated dose, and for unregulated research chemicals no assurance that the powder contains what the label says — analytical surveys of noopept, a compound that does have a dose, found most declared quantities inaccurate.
Isolated compounds
Half-life and dose as each compound’s own monograph records them. These are the amounts used in research or reported in self-directed use, not recommendations.
| Compound | Approx. half-life | Studied dose |
|---|---|---|
| 9-Me-BC Rodent work only, some of it dosed directly into the brain | Not characterised in humans | No established human dose |
| Acetyl-L-carnitine Split 2–3×; absorption is saturable | No simple plasma half-life; body carnitine turnover ~38–119 h | 1.5–3 g/day |
| Alpha-GPC Choline-dense; count it toward the 3,500 mg/day choline upper limit | Not established for the intact molecule | 300–600 mg/day |
| Aniracetam Fat-soluble; split and take with food | Parent ~0.5–1 h; active metabolites ~1–2.5 h | 1,000–1,500 mg/day |
| Bromantane Cumulative — the benefit appeared around day 3 in the large trial | Not established | 50–100 mg/day |
| Caffeine ~400 mg/day is the usual ceiling; 200 mg/day or less in pregnancy | ~5 h (range 3–7 h) | 50–200 mg per dose; 3–6 mg/kg before exercise |
| Centrophenoxine Alerting — take earlier in the day | Short; a prodrug, parent not detected in plasma | 1–2 g/day in the older trials; ~250–500 mg 1–3×/day in nootropic use |
| Citicoline ~21% choline by weight, so label weight is not choline weight | Biphasic; components clear over ~24–56 h | 250–500 mg/day |
| Coluracetam The one figure on record is from a discontinued depression trial | Not reliably established | No validated human dose |
| Creatine Optional 20 g/day loading for 5–7 days; any time of day | Not a classic half-life; stores saturate over days to weeks | 3–5 g/day |
| Dihexa No human has received it in a published study | Not characterised in humans | No established human dose |
| DMAE Labels sit at 100–300 mg/day by convention; the cognitive trials were negative | Not well characterised | No evidence-based oral dose |
| Fasoracetam The ADHD trials titrated to 400 mg twice daily and missed their endpoints | ~1–2 h in animals; not established in humans | No established supplement dose |
| FGL Phase 1 intranasal tolerability doses only | Not established | No established human dose |
| Huperzine-A Micrograms, not milligrams; often cycled | ~5–14 h | 50–200 mcg/day |
| L-theanine Per occasion, 30–60 min ahead; it does not build up | ~1 h | 100–400 mg per dose, commonly 200 mg |
| L-tyrosine Typical 500–2,000 mg labels sit below the studied doses | ~2–3 h | 100–150 mg/kg (≈7–15 g) 1–2 h before a stressor |
| Modafinil Prescription-only; 400 mg adds side effects rather than benefit | ~13–15 h | 200 mg once daily, morning |
| N-Acetyl Selank Amidate No study has administered this molecule | Never measured | No established dose |
| N-Acetyl Semax Amidate No study has administered this molecule | Never measured | No established dose |
| Nefiracetam Supervised trial dosing; food delays but does not reduce absorption | ~3–5 h parent; metabolites ~8–22 h | 600–900 mg/day (research protocol) |
| Noopept Around a hundred times piracetam’s potency by weight | ~15–25 min parent; activity carried by a metabolite | 10–30 mg/day, split |
| Oxiracetam Split 2–3×; reduce the dose in kidney disease | ~3–6 h | 750–1,600 mg/day |
| P21 Every dosing figure comes from rodent chow | Not established | No established human dose |
| Phenylpiracetam Stimulant tolerance builds fast; not late in the day | ~3–5 h (community-reported, poorly documented) | 100–200 mg per dose |
| Phosphatidylserine Take with food; trial benefits emerged over 6–12 weeks | Short in plasma; the effect is membrane incorporation | 100 mg ×3 (300 mg/day) |
| Piracetam Grams, not milligrams; cleared by the kidneys | ~4–5 h | 1,600–4,800 mg/day, split 2–3× |
| Pramiracetam Fat-soluble; take with dietary fat | ~4.5–6.5 h | ~600 mg 2×/day, or 400 mg 3×/day |
| PRL-8-53 One 1978 study, no dose-response, no repeat dosing | Not reported | 5 mg single oral dose — the only human study |
| Racetams The family overview; potency differs by orders of magnitude | Varies across the family | Compound-dependent |
| Sarcosine Every efficacy trial ran inside psychiatric care | Not characterised | 2 g/day |
| Selank Intranasal in the Russian studies; oral delivery is questionable | Short (minutes) | No established Western dose |
| Semax Intranasal in the Russian studies; oral delivery is questionable | Short (minutes) | No established Western dose |
| Sulbutiamine Fat-soluble; divided daily dosing in the trials | ~5 h | 400–600 mg/day |
| Sunifiram Rodent doses only, and they followed an inverted-U | Unknown | No established human dose |
| Uridine monophosphate Commercial convention — no controlled trial of UMP alone at this range | Not established; a precursor pool rather than a drug | 150–300 mg/day |
| Vinpocetine Absorption is ~7% fasting and 60–80% with a meal | ~1–2 h | 15–60 mg/day, split |
Nothing matches.
Nootropic herbs
Herbs are listed separately because their doses are not in the same units and cannot be compared with the table above. A whole plant is a mixture, and it is dosed as a weight of dried herb, a volume of liquid extract, or a standardised extract titrated to a marker compound — not as milligrams of one molecule. Reading “2–4 mL of tincture” next to “300 mg of alpha-GPC” as though they were two points on one scale gets you nowhere.
Liquid-extract doses follow the herbalist’s convention of a weekly volume at a stated extract strength, which is then divided across the week. The liquid herb dosage chart and the herbal tea dosage chart give the full ranges and the arithmetic.
| Herb | Dose |
|---|---|
| Ashwagandha | 1:2 liquid extract, 35–90 mL/week |
| Bacopa The memory effect builds over weeks | 1:2 liquid extract, 30–60 mL/week |
| Black Ant No human dose established; the research used a concentrated fraction in rodents | Powder, capsules or liquid extract |
| Chinese Clubmoss Dosed as the isolated alkaloid, never as crude herb | Concentrated huperzine-A extract, 50–200 mcg/day |
| Codonopsis | 1:2 liquid extract, 30–60 mL/week |
| Coffee Regular intake, not sporadic | 2–6 cups of brewed coffee daily |
| Coleus | 1:1 liquid extract, 40–90 mL/week |
| Fennel | 1:2 liquid extract, 20–40 mL/week |
| Ginkgo | Standardised leaf extract (EGb 761), 120–240 mg/day in 2–3 doses |
| American Ginseng The trial dose | Standardised extract ~2,000 mg/day |
| Asian Ginseng | 1:2 liquid extract, 5–40 mL/week |
| Gotu Kola | 1:2 liquid extract, 15–40 mL/week |
| Guarana Delivers roughly 40–120 mg caffeine; count it toward your daily total | Powdered seed or extract, ~1–3 g/day |
| Guayusa Traditionally a morning leaf infusion of no recorded weight | No codified dose; trials use 600–1,200 mg of standardised extract |
| Holy Basil | Standardised extract, ~1,200 mg actives/day |
| Hops Sedating — an evening herb | 1:2 liquid extract, 10–30 mL/week |
| Juniper | 1:2 liquid extract, 10–20 mL/week |
| Kanna | 100–1,000 mg dried, method-dependent; Zembrin extract 25–50 mg/day |
| Lesser Periwinkle The alkaloid, not the herb | Vincamine 30 mg twice daily |
| Muira Puama | 1:2 liquid extract, 10–25 mL/week |
| Periwinkle The cognitive use is isolated vincamine at 15–60 mg/day, or vinpocetine | Not used as a self-administered simple |
| Rhodiola | 1:2 liquid extract, 20–40 mL/week |
| Rosemary | 1:2 liquid extract, 15–30 mL/week |
| Saffron | Dried stigma ~30 mg/day, or standardised extract ~28 mg/day |
| Sage Culinary amounts; the cognition trials used proprietary extracts | 1 tsp dried leaf per cup as an infusion |
| Samambaia | Standardised extract, 240–480 mg/day |
| Tea The source of both caffeine and L-theanine | 2–3 cups daily, or green tea extract at ~250–500 mg/day of catechins |
| Wormwood | 1:5 liquid extract, 5–20 mL/week |
Nothing matches.
Two herbs in that list are really delivery vehicles for a single molecule. Chinese clubmoss contains so little huperzine-A — around 80 micrograms per gram of dried plant — that reaching a studied dose from crude herb would take several grams a day, which is why the extract is what gets used. Lesser periwinkle is dosed as vincamine for the same reason. In both cases the milligram figure belongs in the compound table, and the plant is the source rather than the preparation.
Stacking
Two combinations are well enough documented to describe as pairs rather than experiments.
Caffeine and L-theanine
L-theanine 200 mgCaffeine 100Roughly 2 parts L-theanine to 1 part caffeine — 200 mg and 100 mg is the usual starting point. Attention trials found the pair sharpened attention while the theanine blunted the jitteriness. Both are short-acting, so this is a per-occasion dose rather than a daily level, and the caffeine half of it still counts toward your daily ceiling and your afternoon cutoff. A cup of tea is the same pairing at a much lower dose.
A racetam and a choline source
Community practicePiracetam, aniracetam, oxiracetam and pramiracetam increase acetylcholine turnover, and the dull frontal headache that regular users report is generally read as the substrate running short. Pairing them with alpha-GPC or citicoline is the standard adjustment. This is community practice rather than a trial finding — no study cited on any of those pages tests the combination — but the mechanism is straightforward and the choline sources are well tolerated.
If you are stacking choline from more than one place, keep the total in view. Alpha-GPC is about 40% choline by weight and citicoline about 21%, and the Tolerable Upper Intake Level for choline is 3,500 mg/day from all sources, eggs included.
Beyond those two, add one thing at a time. Two calming compounds can over-sedate and two stimulating ones can tip into anxiety, and a stack assembled all at once tells you nothing about which part is responsible for what you are feeling.
Tolerance and cycling
This is where most of the disappointment lives.
Stimulants lose their edge fastest
Use when neededCaffeine tolerance builds within days, and much of what a habitual coffee drinker feels in the morning is the reversal of overnight withdrawal rather than a gain over baseline. Phenylpiracetam is reported to lose its stimulant effect quickly enough that continuous daily use is self-defeating; in the trials it was given in courses, not indefinitely. Modafinil is a prescription medicine with its own tolerance profile and its own withdrawal from sleep debt.
The honest use for a stimulant is the day you genuinely need it. Taken daily it becomes a maintenance dose you cannot easily stop, at a cost to sleep that outweighs the alertness it buys.
Enzyme inhibitors are a different problem
Commonly cycledHuperzine-A inhibits acetylcholinesterase rather than supplying a nutrient, and users commonly cycle it — several days on, several off — to limit cholinergic tolerance and side effects. That is a practical convention rather than a trial-established schedule, but the reasoning is sound.
Some compounds are the opposite case and should not be cycled
Keep takingCreatine works by keeping stores saturated, so stopping simply lets them drain over weeks. Bacopa and phosphatidylserine took 6–12 weeks to show a benefit in trials, which means a two-week trial of either tells you nothing. Bromantane built its effect from around day 3 and held it for about a month after dosing stopped.
Deciding whether to cycle comes down to whether the compound acts on a receptor or fills a pool.
Receptors adapt. Pools do not.
A few dosing principles
Keep cholinergics in balance. Racetams draw on acetylcholine; a choline source alongside them often prevents the dull racetam headache.
Stacks are not simply additive. Add one new thing at a time so you can tell what is doing what.
Mind food and timing. Fat-soluble compounds absorb better with a meal; L-tyrosine works better taken away from protein.
Bodyweight is a rough guide, not a rule. Most doses are given as flat amounts because response depends more on metabolism and receptor sensitivity than on size. L-tyrosine is the exception, and it is dosed per kilogram in every trial that found an effect.
Check what kind of compound it is before you check the dose. Some of these are ordinary supplements, some are prescription medicines, and some are unapproved research chemicals with no regulated manufacture. The nootropics index lists the status of each one.
A short trial of a slow compound proves nothing. Match how long you test something to how long its trials ran.
Educational information only, and not a substitute for medical advice. Some compounds listed here (modafinil among them) are prescription medicines in most countries, and several are unapproved research chemicals.
