Andrew Huberman Protocols: 10 Daily Steps to Follow

Cortisol is not the enemy in Andrew Huberman protocols. High cortisol in the first hour after waking is treated as desirable, because it produces low cortisol and higher melatonin later. That inversion runs against most wellness advice about keeping cortisol low.

Huberman laid out the ten steps in a 2025 conversation on The Diary Of A CEO, timed to the release of his book Protocols, and the through-line is sequence: each step changes the conditions for the next one.

The 10 protocols in order, from waking to sleep

The ten protocols Huberman groups in Protocols, his 2025 book with a UK and US edition, map onto a single day. Hydrate on waking, get morning sunlight, train in the first three hours, use a five-minute stress tool, view late-afternoon daylight, eat for stable fuel, darken the sleeping environment, do a bout of hard mental work, keep one daily space with no role to fulfil, and sleep on your side. He says the order matters because the steps interact. How well you slept shapes how the next day feels, and each gear turns the others.

The full set is short enough to hold in your head, and none of it requires equipment beyond water, daylight, and shoes. Huberman's own framing is that the ten were chosen because they move the needle most for mental health, physical health and performance, and that they cost nothing or fit an existing budget.

1. Hydrate with 16 to 32 ounces of water on waking.
2. View low-angle sunlight, or a 10,000 lux lamp, in the first hour.
3. Exercise, resistance or cardiovascular, in the first three to four hours.
4. Use a five-minute breathing practice, or a single physiological sigh in the moment.
5. View daylight again in the late afternoon to buffer evening screen light.
6. Eat complete protein, fibre, fermented food and starches without excess fat.
7. Keep the sleeping environment dark and free of bright LED light.
8. Do 90 minutes of effortful work with the phone out of the room.
9. Take one waking stretch with no role to fulfil.
10. Sleep on your side with your head slightly elevated.

The one item he skips weekly is deliberate cold exposure, which he takes off one day out of seven. Supplements sit outside the ten, and he describes them as an add-on rather than the base.

Why the order follows the cortisol curve

The ten steps follow one biological sequence: a high cortisol peak in the first hour after waking produces the low evening cortisol and strong nighttime melatonin that make sleep come easily. Huberman draws this as a line across a waking day, with the morning peak on the left and the evening trough on the right. When that peak is flat, he argues, every later stressor rides on a higher baseline and lasts longer.

He states the mechanism plainly: the cortisol awakening response is entrainable, which is why people sometimes wake a minute before an alarm. Bright light in the first hour, ideally within 30 minutes of waking and still useful within 60, is what sets that peak. A second pathway runs from the suprachiasmatic nucleus, the master clock in the hypothalamus, down the spinal column to the adrenal glands.

On his diagram there are two more traces alongside cortisol. Melatonin is low in the morning if you slept the night before, and adenosine, a downstream product of ATP use in cells, builds through the day and produces sleep pressure. The healthy curve itself was not invented: it was measured through repeated blood draws across the 24-hour cycle.

His framing of cortisol is deliberate. He calls it an energy-mobilising hormone and a preparation hormone rather than a stress hormone, and points out that the rare condition of chronically excessive production, Cushing syndrome, is not what most people have. Most people, in his account, run a morning cortisol that is too low and an evening cortisol that is too high. A flattened morning curve, he says, is associated with shorter lifespan and slower recovery from illness.

He attaches a prediction to this: 2027 is the year he expects the idea that cortisol is simply bad to be retired. Treat that as his forecast, not a finding. He is also specific that it does not matter whether the morning spike arrives as light, exercise or a stressful email; what matters is the inflection, not the source.

Light timing: bright days, dark nights, and the study behind it

Morning light and evening darkness work as two halves of one signal, and Huberman treats them as additive rather than interchangeable. His morning target is low-angle sunlight, which carries more long-wavelength light and less ultraviolet than overhead sun, so it is easier to look at without squinting or burning. On overcast days the fallback is daylight, since outdoor light on a grey December morning is still far brighter than a lit room, and a 10,000 lux lamp is the second option.

The biology runs through melanopsin-containing intrinsically photosensitive retinal ganglion cells. These cells extend axons into the hypothalamus, where the suprachiasmatic nucleus sits, and low-solar-angle light is the optimal stimulus for them. Huberman's practical version: 10 to 20 minutes outside on a clear morning, longer on grey days, with the caveat that you should not try to do this through a window or a windshield.

Screens do not substitute in the morning. His point is that a phone or laptop is far too dim to drive the cortisol peak, yet bright enough at night to push cortisol up when the body is primed to let it fall. Two effects follow from evening light in his account: melatonin suppression, which is widely discussed, and a rise in cortisol, which he says gets far less attention.

He cites the UK Biobank as the large dataset behind the bright day, dark night framing, and reports that people with bright nights show higher risk of early death and worse metabolic markers. Treat the direction of that association as reported by him rather than a number you can act on directly.

Adolescents are the exception he flags. If you are an adult wondering whether evening daylight buffers you against a late Netflix session, he says it can. He says it does not appear to buffer adolescents in the same way, and that this is part of why a generation of children on tablets at night worries him more than adult screen use does.

Exercise: what the weekly structure looks like

The exercise protocol is a split across the week, not a daily max effort. Huberman describes lifting on three days and cardio on three, alternating, with two to three work sets per exercise and a focus on compound movements that cross multiple joints and muscle groups, such as pull-ups, rows, dips, squats and hip hinges. Cardio splits into one long slow session of roughly an hour, one moderate session of about 30 minutes, and one short high-intensity session built from 30-second efforts with 10-second recoveries on an air bike. The day after a hard leg session is rest.

Timing matters more than he expected. Training in the first three hours of the day produces the cortisol inflection he wants, and after a few days at the same hour the brain begins anticipating the session, which shows up as energy in the 15 minutes beforehand. He compares that anticipation to the hunger signal that arrives before a habitual mealtime. If your schedule forces evening training, he suggests long exhale breathing afterwards and a hot shower, which lowers core temperature through the medial preoptic area and helps sleep.

He separates deliberate cold exposure from the exercise stress response. In his telling, cold water produces trivial increases in cortisol and large increases in adrenaline, noradrenaline and dopamine, which is why he places it in the day for alertness rather than as a cortisol stressor. His stated reason to do it at all is carryover: practising calm while adrenaline is high transfers to unavoidable stress later. On safety, the caveat is his own: cold water is tolerable but not so cold that you hyperventilate deliberately beforehand.

That carryover claim is the reason cold exposure stays on his list at all, even though it is the protocol he skips once a week.

He is not an exercise physiologist, and says so, noting only that he has lifted since he was 16 and was 50 at the time of recording. The weekly structure is his personal template, not a study result.

The 5-minute stress tool and the long exhale

The fastest stress tool in the protocol is one breath: a double inhale through the nose, then a long exhale through the mouth, repeated until the lungs empty. Huberman describes this as the physiological sigh and says a single repetition can take the edge off acute stress, whether before a talk, a diagnosis or a difficult conversation. It costs nothing and needs no equipment.

The double inhale is not decorative. According to work from Jack Feldman's lab at UCLA, a brain area called the parabrachial nucleus generates the double-inhale-then-exhale pattern, and the second inhale reopens the small air sacs in the lungs so the following exhale offloads the most carbon dioxide.

He connects the long exhale to respiratory sinus arrhythmia, the natural slowing of heart rate on exhalation, which runs through the vagus nerve. He is careful about the popular version of vagus talk: about 85% of vagal traffic is ascending sensory information from the organs, and stimulating the vagus in clinical settings can make people more alert rather than calm. What matters here is the descending branch that slows the heart when you exhale.

A randomised study from his own lab compared four conditions over five minutes a day in roughly a hundred subjects wearing heart-rate and sleep trackers: meditation, box breathing, cyclic hyperventilation, and repeated physiological sighs. All four groups improved, and he reports that the physiological sigh group improved most. Melissa Yilmaz Balban built the stress lab and ran the work, with David Spiegel as a collaborator.

For waking in the middle of the night, he pairs ten long exhales with an eye-movement sequence: with eyes closed, move them slowly to each side, then up and down, then in circles each way, then toward the nostrils, then exhale long. The claim is that this pattern quiets the proprioceptive signals from the cerebellum that keep you tracking where your limbs are, which is one of the things that keeps you awake. He says this generally works, and that it may not if you are very stressed.

Nutrition: protein ratio, fibre, fermented food and starches at night

The nutrition protocol is a short list, not a diet brand: enough calories without excess, high complete protein per calorie, vegetables and fruit for micronutrients and fibre, low-sugar fermented food such as sauerkraut or kimchi, olive oil as the main fat, and starches positioned so they help sleep rather than hurt it. Huberman names Justin Sonnenburg's work at Stanford on fermented food and the inflammasome as the clearest evidence in that section.

His protein criterion is his own formulation. He calls it high complete protein to calorie ratio, and gives beans and rice as the counterexample: reaching 30 grams of complete protein from that combination takes roughly four to five times the calories of a chicken breast. He concedes that some people who remove starches or vegetables entirely see autoimmune symptoms resolve, and leaves room for that. On the wider nutrition literature he is blunt, estimating that around 90% of the studies are poor, which is why he does not build a protocol on them.

Fibre needs a caveat he says is unpopular. When Sonnenburg's team had people eat more fibre, inflammatory markers rose in some of them, so more fibre is not automatically better. He also declines the seed oil debate directly, calling the underlying studies too confounded to resolve, on the grounds that comparisons tend to pit a processed diet with seed oils against processed meat.

Starches get a specific role. Foods that raise glucose lower cortisol, which is useful at night when you want cortisol low. He warns against overdoing starches combined with large amounts of fat, and against eating so close to bedtime that sleep suffers, while noting that arriving at bed with gnawing hunger also disrupts sleep.

Focus, hard work and where the phone should be

The focus protocol is 90 minutes to two hours of effortful work, and the phone belongs in another room. Huberman cites a study that put people's phones face down on the desk, in a bag under the chair, or in a separate room. Focus performance was similar across all three conditions, but cognitive flexibility with the material was reduced whenever the phone was anywhere in the room. His reading is that some part of the brain keeps a channel open for the device even when it is silent, so focus costs more energy than it should.

The learning mechanism he describes is error-driven. Reaching for a pen you can already grasp does not change the brain, and he calls the idea that every experience rewires you a public-education myth. Plasticity comes from genuine attempts that fail some of the time, with the failure signal relayed through the cerebellum. He cites a figure of roughly 15% error rate from computational modelling as the productive difficulty range. Drilling a skill you already have, by contrast, aims at 100% performance.

The practical sequence is alert, focused, error-driven, then stop and return a day or two later. Reinforcement through self-testing is what holds the learning in place. He uses large language models for this, asking for a short set of graduate-level questions on a topic and looking up whatever he got wrong. In the studies he cites, students who read a passage once and then self-tested remembered more six months and two years later than students who read it three or four times or highlighted it.

Separately, he flags that the anterior midcingulate cortex enlarges with voluntary difficulty, whether that is dieting, a new exercise programme, or learning to dance if dancing is uncomfortable. He is explicit that he did not discover the structure, only that he discusses its plasticity.

What the supplement and peptide talk does not settle

Supplements sit outside the ten protocols, and Huberman frames them as additions to a base of behaviour rather than substitutes. Asked to keep only five, he names magnesium threonate or magnesium bisglycinate for sleep, at least one gram of EPA omega-3 daily, 5,000 to 10,000 IU of vitamin D, alpha-GPC at 600 to 900 milligrams, and creatine at 5 to 10 grams. Most of these are low-risk and inexpensive; the vitamin D dose sits above common maintenance levels and he notes that people with existing conditions should not extrapolate from his own blood work.

A few adjacent details fill out the list. Prescription EPA such as Lovaza is concentrated and tested for heavy metals. Alpha-GPC raises acetylcholine and supports focus, and he pairs it with a garlic capsule at 600 milligrams because it can raise TMAO, a marker tied to cardiovascular risk. He has taken creatine since he was 17 or 18 and finds it works best for cognitive benefit when sleep-deprived or stressed. Coffee and yerba mate stay: 200 to 400 milligrams of caffeine a day is associated with lower dementia risk, though he calls that correlation rather than causation. About 90% of the world consumes caffeine daily.

He is explicit that the book covers none of the frontier material. Protocols, in his description, is about what he knows for sure, and hormone therapies, peptide therapies and psychedelics are deliberately left out. His personal experiments with peptides are described as self-experimentation by an adult, not recommendations.

One claim in this section deserves caution. He says that pinealon produced large amounts of REM sleep in a month-long trial he ran on himself, taking it every third night. That is a single person's report with no published trial attached in the transcript, and it should be read that way. On other peptides he is equally informal: growth hormone secretagogues gave him deep sleep but destroyed his REM, and BPC-157, a body-protection compound made in the gut, appears in animal data to speed healing of tendons, nerves and ligaments, with the theoretical worry that it could vascularise tumours.

The GLP-1 discussion carries the same caveat. He relays a friend's account of low-dose tirzepatide improving energy and inflammatory markers, and notes that these drugs raise GLP-1 levels far above anything the human body produces naturally. Tirzepatide acts on both GLP-1 and GIP pathways, and Lilly has run large trials in it. On the vision question he pushes back on circulating claims, saying that cases cluster in people with a specific structural feature of the optic nerve head, and that an eye pressure check is a reasonable step rather than a reason to avoid treatment.

Comparison: the ten protocols at a glance

#	Protocol	Timing	Core mechanism
1	Hydration	On waking	Vagal ascending pathway; supports the cortisol peak
2	Morning light	Within 30-60 min of waking	Melanopsin cells and the suprachiasmatic nucleus set the peak
3	Exercise	First 3-4 hours	Big cortisol inflection plus clock entrainment
4	Physiological sigh	Any time	Long exhale slows heart rate via respiratory sinus arrhythmia
5	Late daylight	Last third of the day	Buffers evening screen light in adults
6	Nutrition	Across the day, starches later	Protein ratio, fibre, fermented food; starches buffer cortisol
7	Dark sleep environment	At night	Keeps cortisol low and protects melatonin
8	Effortful work	90 min to 2 hours	Error-driven plasticity, phone out of the room
9	No-role time	Once daily	Touches the unconscious and heart's desires
10	Side sleeping	At night	Glymphatic clearance of brain waste

A typical protocol day, hour by hour

Huberman's own morning runs in a fixed order, and it is worth seeing it assembled before the individual pieces.

1. Wake and lie still for 10 breaths of meditation.
2. Get up, urinate, then hydrate with 16 to 32 ounces of water.
3. Get outside for sunlight on the eyes.
4. Walk the dog, then come back for caffeine.
5. Train in the first three to four hours of the day.
6. Work in a 90-minute to two-hour block with the phone in another room.
7. Get daylight again in the late afternoon.
8. Dim the lights in the last 30 minutes before sleep and get off screens.
9. Do the night-time eye-movement sequence if you wake at 2 or 3 a.m.
10. Sleep on your side with the head slightly elevated.

Night shift workers get their own section in the book. Huberman's warning there is that the studies behind the label cover people on devices after 6 p.m., not only people who work all night, so a late-evening screen habit puts a normal schedule in the same category.

What the ten protocols do not cover

The ten protocols are the behavioural floor, not a full health system, and Huberman says so repeatedly. Three gaps are worth naming because they are easy to over-read.

• Supplements, peptides and hormone therapies are excluded by design. The book is the material he is confident in.
• Sleep duration is treated as a target in its own right. He notes that regular meditators may need about an hour less sleep because short meditation sessions appear to increase glymphatic clearance during the day.
• Walking gets separate treatment from exercise. His target is 7,000 to 8,000 steps a day, not 10,000, and he points to the soleus muscle, about 1% of total musculature, as an outsized glucose sink. Five to ten minutes of walking after a meal flattens glucose peaks, and standing up for air squats every 45 minutes compares reasonably in the data he cites.

FAQ

• Do you have to do all 10 protocols every day? Huberman says he does, with one exception he takes off one day a week. The framing throughout is that the steps compound rather than stand alone, so partial adoption still changes the sequence.

• How much water does the hydration protocol mean? The morning target is 16 to 32 ounces of plain water, with electrolytes optional, and roughly 80 ounces across the day. The emphasis is front-loading fluid and tapering intake toward evening.

• Is morning cortisol supposed to be high? In this framework, yes. A high peak in the first hour after waking is treated as the setup for low evening cortisol and better sleep, and a flat morning curve is the pattern he associates with poor stress recovery.

• Does evening daylight cancel out screen light? For adults, he says a late-afternoon daylight exposure buffers some of the metabolic and sleep effects of screens at night. He says it does not appear to work that way for adolescents.

• Do you need a red light device? No. His position is that sunlight supplies the full spectrum, that long-wavelength light reaches tissue through clothing, and that a device is a fallback rather than a requirement.

• How many minutes of sunlight do you actually need in the morning? Ten to twenty minutes is his rough guide on a clear morning, longer on overcast days. He also says to double up the next day if you miss a day.

• Is cold exposure part of the cortisol protocol? No. He puts cold water in the day for adrenaline and alertness, not as a cortisol stressor, and it is the step he skips once a week.

• What is the actual error rate for productive learning? Around 15%, based on computational modelling he cites. Practising something you already know is different, and there the aim is 100% performance.

• Does the book cover peptides, hormones or psychedelics? No. He deliberately left them out because he treats them as the frontier rather than settled ground.

Building this into your own week

The ten protocols are less a checklist than a set of anchors: light in the morning, movement early, one hard block of mental work, a deliberate pause, and a dark, cool room at night. Trying to install all ten at once is the failure mode, since the steps depend on each other and the sleep they rest on is the last one you fix.

If you start anywhere, start with the first two. Hydration and morning light are free, take minutes, and set the cortisol peak that every later step leans on. Add the training block once the mornings are stable, then the phone-out-of-the-room work block, then the evening dimming. Treat Huberman's forecasts, including the 2027 cortisol prediction and his personal peptide experiments, as opinions rather than findings.

For anyone who wants to keep going deeper, the full conversation is the primary source, and most of the sections above map directly onto a chapter of Protocols.

Source video

Turn your own video into a written guide

If any part of this routine stuck, you already know the shape of the problem it solves: a long, dense conversation contains real value, and almost none of it survives outside the video it came from. The same is true of your own recordings. A talk, an interview, a walkthrough, a lesson you explained out loud once and never wrote down.

Skalablog takes a YouTube URL, transcribes the video, and turns it into a structured article you can edit and publish. If you have expertise sitting in a video that only ever gets watched once, that is the content worth rescuing.

Skala Blog