I use the Skin Deva brand and so far (it's been two weeks) can see an improvement. I bought it directly from their website
PositiveTopical · forehead lines
Skin Deva serum, in reply to a request for Leuphasyl products; two weeks in
Users report
Too few reports to judge. Leuphasyl is hard to buy on its own (a user pointed to Skin Deva, and DIY suppliers sell the raw solution), and most people meet it unknowingly inside multi-peptide 'Botox-alternative' serums next to argireline, Matrixyl and SYN-AKE, where nobody can credit it. One user of a Skin Deva serum saw an improvement in forehead lines after two weeks; another, who says they are biased toward the compounding pharmacy they buy from, credits it with GHK-Cu for smoother, tighter skin. A German microneedled treatment (NoTox) combining it with argireline drew questions but no outcome reports. The score is closer to a placeholder than a verdict. Felt nothing: Too few reports to count (2, one promotional).
Best effects
Claimed: softer expression lines on forehead and around the eyes · one user: visible improvement within two weeks
Downsides
Hard to find as a single-ingredient product · no meaningful user record
Evidence
Small human studies — one uncontrolled 20-volunteer study (2014, Romania) found 2D wrinkle traces shrank about 35% between the brows and 28% around the eyes with a 2% cream over two months, and a 2024 Polish study used it with retinol in lipid nanocarriers; otherwise supplier, cell and computer-model data. No placebo-controlled trial of the peptide alone
Caveat
Pentapeptide-18 is [D-Ala2]-leucine-enkephalin, an opioid-receptor agonist long used in pharmacology research; the cosmetic story is that it damps acetylcholine release at the nerve ending. At cosmetic levels on intact skin there is no evidence it reaches nerves or the circulation, and peptides of this kind do not cross into the brain, so no opioid-like effect is expected. The 2014 paper miscalls it 'pentapeptide-3'. This verdict covers creams; microneedled 'NoTox' treatments put it deeper and have no published safety data.
Typical use
2% of the Leuphasyl solution was the effective level in the one human study; finished serums rarely say. Apply over the muscles that make the lines (between the brows, crow's feet) once or twice daily.
Getting it
Over-the-counter cosmetic: a Skin Deva serum one user bought for it, the raw Lipotec solution from DIY suppliers (Lotioncrafter), and multi-peptide serums such as DNA Code's and Meder's Myo-fix alongside argireline; in Germany some clinics microneedle it in as 'NoTox'.
Stacked with
Argireline: The supplier designed it as argireline's partner: argireline at SNARE assembly, Leuphasyl upstream at the enkephalin receptor. · GHK-Cu (topical): One user's pairing.
Yes I use them and they work. They definitely smooth the skin and make it look younger and stronger/tighter
PositiveTopical · small wrinkle forming · vendor-adjacent
uses GHK-Cu and Leuphasyl together from a compounding pharmacy they say they are 'biased' toward; earlier pasted a chatbot's description instead of their own results
Safety
Tier by safety barStrict CautionBalanced CleanEvidence only Clean
Seen harmMechanismData gapSilent harmProduct
Seen harm 0/3
No serious adverse events in the 20-volunteer study, the 2024 nanocarrier study or first-hand reports; exposure is modest and mostly hidden inside multi-peptide serums, so the zero is weak reassurance. Dragomirescu et al., Cosmetics 2014 (n=20, two months) ↗
Mechanism 1/3
Chemically [D-Ala2]-leucine-enkephalin, an opioid-receptor agonist; a systemic or central opioid effect would need it to cross intact skin and the blood-brain barrier, which a hydrophilic pentapeptide used at a fraction of a percent of a dilute solution is not shown to do. Speculative, hence 1. Lipotec Leuphasyl technical paper (enkephalin-receptor mechanism) ↗
The only plausible harms at cosmetic exposure are irritation or contact allergy, noticed at once and gone on stopping.
Product 0/3
Bought as finished over-the-counter serums; DIY mixing of the raw solution and clinic microneedling ('NoTox') are minority channels, the latter delivering it below the barrier with no safety data.
Never if
known allergy to a product containing it · pregnancy and breastfeeding (no reproductive data; opioid-receptor agonist by design)
Studies 2
Dragomirescu et al., The Efficiency and Safety of Leuphasyl, a Botox-Like Peptide (Cosmetics 2014)open-label · n=20 · 2014 Creams with 0.5%, 1% and 2% Leuphasyl applied over the corrugator and orbicularis muscles for two months: only 2% gave important improvement, with wrinkle-trace size down 34.7% between the brows and 28.4% around the eyes on 2D imaging; no wrinkle disappeared. No control group.
Each bead is one amino acid in the peptide's chain, coloured by its side chain: grey for non-polar, cyan for polar, blue for positively charged and fuchsia for negatively charged. Length and mix tell related peptides apart at a glance, from a two-residue dipeptide like carnosine to a 43-residue protein like thymosin beta-4. The colours describe chemistry only and predict neither effect nor safety.
non-polarpolarpositivenegative
Report grades
Every first-hand Reddit or X report is graded on its own, from glowing through positive, mixed, nothing and negative to bad reaction, and the bar shows how those grades split. Posts that look vendor-adjacent stay on the peptide's page but are left out of the counts.
GlowingPositiveMixedNothingNegativeBad reaction
% positive
The share of a peptide's graded Reddit and X reports that are glowing or positive, vendor-adjacent posts left out. A peptide with few reports is pulled toward the average of all reports, as if it had eight average ones besides its own, so two glowing posts cannot top the list. Kept peptides rank by the same reports' mean grade, shrunk the same way, and excluded ones follow in that order.
Researcher estimate
No first-hand report of this peptide has been graded yet, so the figure is the researchers' own rating out of 10 from what they read about it. It ranks alongside the computed scores until reports arrive.
Safety tier
Each peptide is tiered clean, caution or excluded from its five risk axes under the chosen safety bar. An excluded peptide has a plausible route to serious harm, so it stays listed with the reason but ranks behind every other entry, however good its reviews.
Safety bar
The safety bar sets how much risk it takes to exclude a peptide. Strict: out on any serious mechanism with little human data, or any documented serious harm. Balanced: out on documented serious harm (unless it is obvious, reversible and from a regulated product), or a serious mechanism whose harm would be silent with no human data at all. Evidence only: out only on documented serious harm in humans that is lasting, delayed or from grey-market product.
Seen harm
Documented serious adverse events in humans, from 0 for none despite real exposure to 3 for deaths or several well-documented serious harms. Serious means a hospital or emergency visit or harm lasting weeks, so a headache or a bad night does not count.
Mechanism
A plausible biological route to serious harm at the doses people use, such as growth-factor signalling, hormone-axis suppression, heart or blood-pressure effects or immune activation. It runs from 0 when none is known to 3 when a dangerous pathway has been shown in living animals.
Data gap
How little human safety data exists for the molecule actually sold, from 0 for an approved drug with large trials and years of use to 3 for no human safety data at all. A modified analogue does not inherit its parent's trials.
Silent harm
Whether the feared harm would show quickly or stay hidden, from 0 when it is obvious and goes on stopping to 3 when it is silent and delayed, like tumour promotion or slow organ damage. User reports cannot reveal a harm that takes years to appear, so a high score here limits what good reviews can tell you.
Product
The risk of what people actually buy, from 0 for a regulated pharmacy or supplement product taken by mouth or on the skin to 3 for a grey-market injectable. In one dataset of 6,441 grey-market samples, 42–71% failed basic quality criteria and 15% carried endotoxin.
Evidence
The strongest kind of study behind the peptide's effects, from an approved drug through human controlled trials and small human studies to animal work or nothing of its own. Animal results often fail to carry over to people.
Access
How people get it: as an approved drug, a compounded prescription, an over-the-counter supplement or from the grey market, where nobody checks the vial. The channel decides both the legal position and the risk of a bad batch.
Price per month
Thirty days of the typical dose, from its low end to its high end, priced at the listing a careful buyer would choose: in stock, from a shop scoring 4 of 5 on trust (third-party tested where no score is set), failing that 3, a pure pack before a blend, a European shop first, then the cheapest per mg. The range is the dose's, not the shops'. A per-dose amount counts once a day unless the dose says otherwise. Prices are in euros at ECB rates of 2026-10-29, before shipping; the shop table keeps each listing's price per mg, per gram for bulk supplements. Where no listing can be priced the researchers' own estimate stands in, and says so.
Dose users report
The middle half of the daily doses stated in at least five first-hand reports, vendor posts left out. Where its middle sits more than twice above or below the declared dose, the month is priced at it as well, since what people take is what they pay for.
Testing
Whether the shop publishes lab results: third-party tested means an independent lab checks each batch, own COA means the shop's own certificate of analysis, and no public tests means neither. A certificate shows purity, not safety.
Added peptide
Some packs sold under one peptide's name hold a second one, such as 5 mg GHK-Cu in a nasal kit, stated only in the description. Sprayed in the nose or injected it reaches the blood like an injection, so it carries the risk of that route and is tiered as the injectable form, whatever its tier on the skin. A pack whose page gives no contents list may hold more than its name.