An evidence-graded nootropic repository
The word "nootropic" was invented in 1972 to describe piracetam. Piracetam doesn't work.
Noograde sorts the published research behind each compound by where it sits on the hierarchy of evidence, then reports four things separately: what the evidence concluded, how confident we can be, how big the effect actually is, and what it costs you. We sell nothing. Every other site on this shelf does.
How we grade — three questions, kept apart
a verdict belongs to a claim, not a moleculeIs there human evidence?
Extensive, some, minimal, or none — for any condition. Before asking whether it works, ask whether anyone checked.
Is it safe?
Graded independently of efficacy. A compound can work and still be the worst thing on the shelf. Ask phenibut.
Works for what?
Each indication gets its own grade: Works, Mixed, Doesn't work, Too little data, or Never studied.
"Never studied" is a verdict
Not a maybe, not an open door. If nobody ran the trial, that is the finding — and here, it is usually the finding.
The row that matters
Every entry carries one highlighted row: cognitive enhancement in healthy adults. That is what you came for.
15 compounds, graded per claim
Each entry is graded for one named indication. A compound can be Strong for one use and Refuted for another — the verdict travels with the claim, never with the molecule. Inclusion criterion: anything people take to change how their brain works.
Everything here that works, works by fixing a deficit.
Some of these compounds are real drugs. Modafinil is FDA-approved and it genuinely works. So does caffeine. So does piracetam. Look carefully at what they work for — and at what happens when you ask them to improve a brain that isn't broken.
— of —
Compounds here with a demonstrated benefit for anything at all. Modafinil for excessive sleepiness — FDA-approved, not in dispute. Caffeine for alertness under sleep loss. Nicotine for smoking cessation. Piracetam for cortical myoclonus. Magnesium for magnesium deficiency. Oxytocin, intravenously, for labor induction.
Every one corrects a deficit. Sleep loss. A sleep disorder. Myoclonus. An actual nutrient shortfall. Labor.— of —
Compounds never tested for cognitive enhancement in healthy adults. The — that were tested — caffeine and modafinil — came back mixed or thin, never convincing: modafinil's attention effect is g = 0.06.
Not one compound here is established to enhance a healthy brain.That is the pattern, and it is the whole site in one line: an approval is a statement about one molecule, one route, one indication — and it does not travel. Modafinil cleared real trials for narcolepsy; that is not a licence to read across to your Tuesday. Oxytocin is FDA-approved for labor induction and refuted for autism. The drugs on this shelf that work are correcting something broken. None of them has been shown to turn the dial past normal.
Your brain is a gated community
Almost every claim on this shelf quietly assumes the molecule arrives. For most of them it does not — and you can predict which ones from a single number you can look up in thirty seconds.
What the blood–brain barrier actually is
Everywhere else in your body, capillaries leak on purpose. The cells lining them sit side by side with gaps between, and small molecules drift through into tissue. That is how your muscles get fed.
Brain capillaries are built differently. The endothelial cells are welded together by tight junctions — protein seals that close the gaps entirely. They are wrapped in pericytes and astrocyte end-feet. And they are studded with efflux pumps like P-glycoprotein, which grab molecules that did get in and throw them back out. It is not a wall so much as a customs checkpoint with a bouncer.
One number predicts almost everything on this site
Every compound in the repository, plotted by molecular weight. Colour is our independently assigned CNS exposure grade — how well established it is that the thing reaches a human brain. The two were graded separately, from different evidence. Look what happens anyway.
The pattern is strong — and the exceptions are the lesson
Sort the repository by weight and a clear tendency appears: almost everything with demonstrated human CNS exposure clusters at the light end, and the heavy peptides — 700 to 1,400 daltons, charged, water-loving — are the ones whose brain access is presumed, contested, or never measured. Weight came from the molecular formula; the exposure grade came from reading pharmacokinetic literature. They line up because the barrier does not care what the product page says.
But the fit is not perfect, and the exceptions teach more than a clean line would. Methylene blue is 320 daltons and demonstrably reaches the human brain — you can watch it on fMRI — because it is lipophilic and uncharged, and lipophilicity buys back some weight. Meanwhile dopamine at 153 daltons cannot cross at all, and magnesium is the sharpest case of all: the active species is Mg²⁺ at 24 daltons, the lightest thing in this repository by a wide margin, and it still does not freely enter the brain, because it is charged. The barrier does not admit ions by passive diffusion at any size — and for magnesium it goes further, actively regulating CSF magnesium against blood levels, which is why magnesium L-threonate had to be engineered and why its central claim is still unsettled.
So read the chart as a filter, not a verdict. Above roughly 500 daltons, passive entry is essentially off the table — the peptides fail here outright, and no amount of marketing changes it. Below it, weight only buys a hearing: lipophilicity, charge, transporters and efflux pumps decide the rest. The small molecules pass the filter and are then judged on everything else.
Three entries sit off the chart because they have no single molecular weight: cerebrolysin (a heterogeneous porcine brain extract, different in every batch), magnesium (an ion sold as salts from 40 to 294 Da — a formulation question, not a diffusion one), and microdosing (a practice, not a molecule).
Be small and greasy
Passive diffusion. Under ~500 Da, lipophilic, few hydrogen-bond donors. Every CNS drug you have heard of takes this door — every SSRI, every benzo, ketamine, caffeine. All under 320 Da. It is not a coincidence; it is the entry requirement.
Steal a key
Carrier-mediated transport. Dopamine is 153 Da and cannot cross at all — too polar. So Parkinson's is treated with L-DOPA (197 Da), which is heavier but hitches a ride on the LAT1 amino-acid transporter. L-theanine takes the same door, on the same transporter. Small is necessary. It is not sufficient. You need the right key.
Use the side entrance
A few brain regions have no barrier at all — the circumventricular organs, like the area postrema. This is how GLP-1 drugs suppress appetite without ever crossing. A real door, but it only opens onto a few rooms.
The maximum share of an intranasal oxytocin dose that reaches the cerebrospinal fluid, across the seven CSF studies reviewed by Leng & Ludwig in Biological Psychiatry (2016). Their title: "Myths and Delusions."
So why is everything sold as a nasal spray?
Because the industry knows about the weight problem. Intranasal dosing is the workaround: nerves running from your nose to your brain — olfactory and trigeminal — may ferry molecules past the barrier entirely. That pathway is real. It is also rarely quantified in humans, and it is doing enormous unexamined work on this shelf.
When somebody finally measured it carefully, for the most-studied intranasal peptide in psychiatry, the answer was 0.005% of the dose. Meanwhile blood levels went supraphysiologic — so the drug went somewhere, mostly your gut and heart.
What to do with this: when a product page says "crosses the blood–brain barrier," ask two questions. Measured in whom? — usually the answer is rats, once. And how much? — usually nobody knows. Then look up the molecular weight. If it is over 500 and the answer to both is a shrug, you have learned what you need to know.
One caution, though — this cuts both ways. Getting in is necessary, not sufficient. Piracetam sails across the barrier, modulates AMPA receptors on arrival, and still does nothing measurable for cognition. The barrier explains why most of this shelf cannot work. It never promised that the ones that cross do.
Why "intranasal" is not the shortcut it sounds like
Nose sprays are the workaround for the weight problem — nerves running from the nasal roof to the brain can ferry molecules past the barrier. But for a conventional spray, the direct nose-to-brain fraction is a small slice of a small slice, and almost every clean number in the nootropic literature is borrowed from rodents whose noses are built differently from yours.
Once the plume lands, every molecule faces three fates. Only two of them are actually "direct."
Olfactory nerve
Across the olfactory epithelium in the roof of the cavity, through the cribriform plate, to the bulb and CSF. A fast perineural component reaches CSF in minutes; a slow intraneuronal one takes hours to days.
Trigeminal nerve
Along CN V branches innervating the mucosa, entering at the pons and brainstem. Seeds caudal structures — hindbrain rather than frontal cortex.
Systemic absorption
Across the respiratory mucosa into blood — then it has to re-cross the BBB to reach brain. For a peptide with a 2–5 min half-life that is not BBB-permeant, this is a near dead end: it degrades before it can back-cross.
Where the geometry bites
The olfactory region — the only mucosa with a true direct CNS conduit — is about 10% of the human nasal cavity (~2,097 mm²). And a conventional metered spray throws its plume low and forward, depositing almost nothing that high and far back. Purpose-built olfactory devices aim for ≥50% olfactory-cleft deposition. A drugstore pump is nowhere close.
Here is the compounding cruelty for a peptide: because Route 3 is a peptidase graveyard, the tiny direct fraction ends up carrying essentially the entire central signal — even though in absolute terms it is small. The systemic majority is absorbed, then largely wasted.
(the only direct-to-brain mucosa)
(systemic → mostly wasted for a peptide)
The translational landmine
The flattering Semax numbers — the 60–70% CSF-to-plasma ratio, the brisk uptake — are from rat tritium studies. The rat olfactory epithelium occupies a far larger share of the nasal cavity than the human's ~10%, and rats are often dosed with targeted olfactory deposition while you are using a broad spray. So the rodent direct-transport fraction is a systematic overestimate of what a human gets.
No human DTE/DTP has ever been published for Semax. The human direct fraction is genuinely uncharacterized. Every clean number in the nootropic literature is rodent-derived — including the one that makes intranasal peptides sound like they work.
How a peptide gets graded
Every entry is built the same way, so any two compounds can be compared on the same scale rather than on marketing volume.
Fix the claim before grading it
A verdict is issued for one compound and one named indication. "Does oxytocin work?" is unanswerable; "does intranasal oxytocin improve social functioning in ASD?" has an answer.
Assemble the evidence base
Collect the relevant published studies — in vitro, animal, uncontrolled, and randomized — plus systematic reviews where they exist.
Rank each study by level
Each study is placed on the evidence hierarchy (Oxford CEBM 1–5). A rat model and a 2,000-patient RCT are not weighted equally.
Build the evidence ladder
Studies stack by level, and each rung records what the result said — positive, null, or mixed. Height shows how well the question was tested; the markers show what came back.
Issue verdict and certainty separately
The verdict is the conclusion. The certainty is how much a future trial could move it. Keeping them apart is what lets "Refuted" and "Insufficient" stay distinct.
Grade risk the same way — level, then confidence
Hazard and how well that hazard is known are different facts. A compound nobody has studied will look safe on an adverse-event table. Every risk rating here carries the exposure denominator standing behind it.
Why volume ≠ strength
A compound can have hundreds of papers and still earn a low grade. If they are nearly all animal or cell studies, the base is wide but shallow — it never climbs to the controlled human trials that tell us whether something works in people.
The ladder is built to make that shape visible at a glance: BPC-157's is broad and short, rapastinel's is narrow and tall. Two very different research programmes, immediately legible.
Watch for the tall ladders topped with a red ✗ — oxytocin, cerebrolysin, davunetide. Those are the most rigorously answered questions in the repository, and the answer was no.
Frameworks: GRADE Working Group; OCEBM Levels of Evidence (2011).
Worked example: is intranasal oxytocin safer than caffeine?
On a naive comparison, yes. In SOARS-B, 290 children took intranasal oxytocin for 24 weeks with adverse events indistinguishable from placebo, and no serious events in the Phelan-McDermid trial. No dependence, no tolerance, no withdrawal. Caffeine has two DSM-5 diagnoses attached, documented fatalities at the powdered extreme, and reliably degrades sleep. Oxytocin appears to win.
It doesn't, and the reason is the denominator. Caffeine's risks are known because of billions of person-years. Chronic intranasal oxytocin has roughly 300 subjects, maximum 24 weeks. "Low risk" means characterized and modest for one and unobserved in thin exposure for the other — the same words carrying wildly different confidence.
Two further asymmetries. Oxytocin is partly safe because it barely arrives — at ≤0.005% of dose reaching CSF, giving safety credit for failure to deliver is strange accounting, and caffeine's main risk exists precisely because it reliably works. And intranasal dosing drives peripheral oxytocin supraphysiologic, where receptors sit on myometrium, vasculature, heart and gut; Pitocin's known dose-dependent harms show the molecule is not peripherally inert, and nobody has characterized chronic supraphysiologic exposure outside obstetrics.
So the honest answer is: narrowly yes, broadly no. That answer is only visible because level and confidence are graded apart — which is why they are.