Cutting-Edge Genomics Meets AI Innovation
Our platform is built on a real genetic trait catalogue, a report-reading pipeline that checks its own work, and honest reporting about what genetics can and can't tell us.
5 Trait Categories, 1,020+ Genetic Traits
Our master catalogue organises traits into 26 profiles across five categories. Each family's report typically covers 113-117 of these traits.
Diet & Nutrition
Nutrient TendenciesVitamin and mineral tendencies (such as Vitamin D and B12), how the body handles carbs and fats, taste sensitivity, and snacking tendency.
Academic & Cognitive
Cognitive SkillsCognitive ability, memory, attention and focus, motivation, reading literacy, mathematical ability, verbal ability, creativity, and stress resilience.
Fitness & Sport
Performance TraitsPower-versus-endurance orientation, muscle and recovery traits, injury-risk tendency, and sport suitability.
Personality & Talents
15 Traits + 7 TalentsFifteen personality dimensions, aspects of intelligence and emotional intelligence, and seven distinct talent areas.
Health-related
Clinically ReviewedElevated tendencies toward specific conditions, allergies, sleep traits, and neurological traits. Shown to parents and counsellors with careful clinical framing — risk-related traits are never shown to a child directly.
From Lab Report to Insights
GenExcel doesn't run the lab test — we read the report it produces, automatically and self-verified, before a human ever reviews it.
Lab Report Received
Your certified testing lab completes the genetic analysis and issues a PDF report. GenExcel picks up from there — turning that report into something a family can actually use.
Exact Reading (fast path)
For known lab report formats, our exact reader parses every trait row and gene reading, measures coloured bars, sliders and gauges to recover LOW / MEDIUM / HIGH bands, and detects ticked boxes and colour badges — in about 10 seconds, with no AI interpretation and no per-report cost.
Flexible Reading (AI-vision fallback)
For unfamiliar report layouts, a second AI-vision-based reader takes over — about 8-9 minutes per report, at roughly 93% accuracy on row and value recall and roughly 96% agreement on gene readings. Its known weak spot is reading colour bands on chart pages, which is exactly why the exact reader is used wherever it applies.
Four-Way Self-Check
Every reading is checked against the source report itself: nothing was dropped, nothing was invented, gene readings agree with the report's own glossary, and row counts match what's actually drawn on the page.
Human Review & Delivery
Anything the flexible reader is uncertain about is routed to a human reviewer before it ever reaches a family. Only then is the personalised report generated.
Every Report Passes Four Checks
Before a report reaches a family, it's checked against the lab document itself — automatically, every single time.
Nothing Dropped
Completeness Check
Confirms every trait row in the lab report was captured
Run on every report, every time
Nothing Invented
Fabrication Check
Confirms no value was added beyond what's in the report
Run on every report, every time
Glossary Match
Consistency Check
Confirms gene readings agree with the report's own glossary
Run on every report, every time
Row-Count Match
Rendering Check
Confirms row counts match what is actually drawn on the page
Run on every report, every time
Confirmed Against the Real Product
Our report-reading pipeline was checked against the working product itself, not an early plan or a demo. Here's what it actually does today.
Our Commitment
We never overstate what genetics can tell us. Our reports clearly distinguish between strong and emerging evidence.
See the Science Behind Your Child's Report
From lab report to a tutor that teaches differently — backed by a reading pipeline that checks its own work.