Every claim here is graded by how strong the evidence actually is — and linked to the source so you can check it yourself. Where the science is thin, we say so.
900+
Studies read
29
Oils graded
4
Honest grades
100%
Linked to PubMed
Four grades. No spin.
Tap a grade to see what it means. We never call something proven when it isn’t.
ModerateControlled human trialsRepeated studies in people point the same way.
EarlySmall / lab / animalPromising signals, but not yet proven in people.
TraditionalLong-usedCenturies of use; clinical evidence still anecdotal.
NoneNo trials foundWe looked and found none. We won’t pretend otherwise.
Explore the evidence
Filter by oil, concern or proof strength — every card links to its study.
Oil
Concern
Grade
0 oils · 0 studies shown
Summary of our evidence database (updated 2026-07-16) — use the filters above for the interactive view; every claim links to PubMed on the oil pages.
Traditionally used for acne across parts of Asia — lab tests show its natural compounds fight acne-causing bacteria, though no human trial has confirmed it works on skin yet.
A vitamin-B12 cream containing avocado oil matched a prescription psoriasis drug in a small trial, with fewer side effects — though avocado oil itself wasn’t tested alone.
Castor oil-infused wound dressings showed antibacterial activity in lab tests, and it’s been used clinically to help surgical wounds close — though rigorous controlled trials are still missing.
A rheumatoid arthritis trial found the GLA compound in evening primrose oil eased joint tenderness — though the credit is shared with a related oil and, in blended studies, with fish oil.
A trial found a coconut-oil cream worked as well as a plain glycerin cream for hand eczema — holding its own against the comparison, even without beating it outright.
Lab tests found macadamia oil itself didn’t fight candida yeast — though a related plant compound (not the oil) showed antifungal promise in separate research.
Moringa’s broader anti-inflammatory properties are well documented in reviews, with brief mentions of joint benefits — though no dedicated clinical trial has tested it for joint inflammation specifically yet.
Pumpkin seed oil shows up in reviews of herbal psoriasis therapies, and a lab-model formulation using it as a carrier reduced inflammation — early signals, not yet a human trial.
Two clinical trials on rosehip for joint inflammation disagree — one found modest relief for rheumatoid arthritis, the other found none. We’re showing you both, not just the flattering one.
In animal burn studies, sea buckthorn sped healing, boosted collagen, and grew new blood vessels faster than untreated wounds — human trials haven’t been run yet.
The closest study used sweet almond oil only as a carrier in a calming blend — not tested alone, and not really about wrinkles. We’re not going to dress that up as anti-ageing proof.
113 studies on clove oil and fungal infections — strong lab evidence, including a nanoparticle formulation that beat ringworm fungus in testing. Still waiting on human trials.
Eucalyptus-infused wound dressings sped healing in lab and animal studies. One real caution: a documented case of serious eye injury from a eucalyptus product — keep it away from eyes.
Frankincense lowered stress hormones and improved sleep in animal studies, and aromatherapy massages that included it eased anxiety in cancer patients — though most human evidence comes from blends, not frankincense alone.
A hospital study found aromatherapy blends including lavender significantly eased anxiety and stress in cancer patients — a real result, though tested as part of a blend, not lavender alone.
One small trial found lemon scent didn’t ease pain during a medical procedure the way lavender did in the same study — we’d rather tell you that than dress it up.
113 studies point to oregano oil fighting fungal infections in the lab — including a nail-lacquer formulation that beat nail fungus in realistic testing. Human trials are still to come.
Long used for hair and scalp. Early studies are promising — one small trial even matched 2% minoxidil for regrowth — but the evidence base is still limited.
65 studies back sweet orange for easing anxiety — several human trials in dental chairs and waiting rooms found lower stress and cortisol from just smelling it, though many were small studies.
Tea Tree Oil — Moderate evidence · 22 studies read
In controlled trials a 5% tea-tree gel cut acne lesions about as well as benzoyl peroxide — with less dryness. Eight acne studies make it our best-evidenced oil.
Small human trials found ylang ylang eased tension and lowered blood pressure — mostly tested in blends with other oils, so the effect of ylang ylang alone isn’t fully proven yet.
Study counts reflect our internal research database (900+ graded study citations across the range; 500+ unique papers linked by PMID). Figures are honest tallies, not marketing claims; individual uses link to PubMed.
How we grade
No clever wording. We read what exists, weigh how strong it is, and link the source.
We read the literature
Real papers from PubMed and clinical databases — not blog posts. 900+ study citations across the range and counting.
We weigh the strength
A human trial outranks a mouse study, which outranks a lab dish. The grade reflects that hierarchy — honestly.
We link the source
Every use links to the actual study. If we found nothing, we grade it None — and say so out loud.
Questions, answered straight
Why grade your own products "Early" or "None"?+
Because it’s true. A carrier oil can be excellent without a stack of trials behind it. We’d rather tell you exactly where the evidence stands than inflate a claim and lose your trust.
What counts as "Moderate" evidence?+
Repeated, controlled studies in actual people pointing the same way — like the acne and headache work behind Tea Tree and Peppermint. It’s our highest grade; we don’t use the word "proven."
Where do the study counts come from?+
Our internal research database — an honest tally of the papers we’ve read per oil — 900+ graded citations, 500+ unique papers linked by PMID. They’re a measure of how much we’ve looked, not a marketing number.
Can I read the studies myself?+
Please do. Every individual use links straight to PubMed. Open a card, find the claim that interests you, and click through.
