Scan Your Label, Discover Your High
Your cannabis experience isn't random—it's chemistry. Snap a photo of your label and our AI predicts your high from the terpene profile.
What's In Your Stash?
Upload a photo of your label—we'll extract the terpenes and tell you what kind of high to expect
From Label to Insight in Seconds
Terpenes are the aromatic compounds that shape your high.
1. Snap
Photograph your product label, COA sticker, or dispensary tag.
2. Analyze
Our AI extracts terpenes—the compounds that shape your experience.
3. Discover
Get matched to a High Family so you know what to expect.
Why Terpenes Matter More Than THC%
Ever wondered why two strains with the same THC hit completely different? Terpenes are why. These aromatic compounds—like myrcene (relaxing), limonene (uplifting), and caryophyllene (soothing)—work alongside cannabinoids to create your unique experience. We call this the entourage effect. Our scanner reads these compounds to predict your high.
The 6 High Families: Patterned on Peer-Reviewed Chemistry
"Indica" and "sativa" tell you about the plant. A 2025 German study analyzed 140 cannabis flowers and identified 6 distinct terpene clusters — our six High Families are patterned on them. The study numbered the clusters; the names below are ours.
Citrusy limonene strains for mood elevation and social energy.
Rare terpinolene-and-ocimene strains — just 6 of the study's 140 flowers.
Myrcene-heavy strains for deep body relaxation and sleep.
The least aromatic, most even terpene profile — subtle, not weak.
Caryophyllene strains that target CB2 receptors for therapeutic relief.
Full-spectrum strains where every terpene pulls together — complex and potent.
Why We Ditched "Indica" & "Sativa"
"Indica" and "sativa" describe the plant — not your high. A 2025 German study (Cannabis and Cannabinoid Research, DOI: 10.1089/can.2024.0127) analyzed 140 flowers and found the labels don't predict chemistry (p > 0.05). Instead, the researchers found that 9 terpenes explain 86% of chemical variation and identified 6 distinct clusters using K-Means analysis.
Our six High Families are patterned on those 6 clusters — so a scan sorts your label by measured terpene chemistry rather than by a marketing term. What that chemistry feels like is our reading, not the study's: it clustered the chemistry and left the effects to future clinical trials.
Read the Full BreakdownQuestions? We've Got Answers
Everything you need to know about label scanning
What types of labels can I scan?
The scanner works best with COA (Certificate of Analysis) stickers that show terpene percentages, but it can also extract data from retail packaging labels. Clear photos with good lighting produce the best results.
How accurate is the classification?
Accuracy depends on available data. Full terpene profiles (from COA labels) provide 80-95% accuracy. Labels with only THC/CBD and dominant terpene provide 60-80% accuracy. Basic labels with just strain type give lower confidence but still useful categorization.
Is my label image stored?
No. We do not store your images. The label is processed by AI, classified, and immediately discarded. Your privacy is protected—we never save or share your product photos.
What information is extracted?
The scanner extracts: strain name, brand name, product type, THC/CBD percentages, full terpene profile (if available), dominant terpene, strain type (indica/sativa/hybrid), batch number, lab name, and test date.
What's the science behind High Families?
The chemistry comes from a 2025 peer-reviewed German study of 140 medicinal cannabis flowers. Using K-Means clustering, the researchers identified 6 distinct terpene clusters built on 9 key terpenes (which explain 86% of chemical variation), and found that traditional "indica" and "sativa" labels have no statistical correlation to chemical composition (p > 0.05). The study labels those clusters by number and names none of them — the High Family names, and the effects we attach to them, are our reading of the chemistry rather than study findings. Read more →
Your images are processed instantly and never stored.