Understand the evidence behind your tools
GC/MS ANALYSIS
Compare batches in context
The workspace compares reported constituent percentages with earlier batches from your report library. It shows observed differences and the records behind them.
- Match the analytical context. Reports must share the botanical species, reported chemotype, plant part, extraction, analytical method and measurement basis. Origin matching is optional. Unknown identity fields prevent a historical match.
- Build a historical reference. Each reference uses strictly earlier analysis dates. Selected reports, repeated source documents and duplicate batches are excluded. A finding requires at least three quantified observations of that constituent.
- Investigate departures. Values above or below the observed minimum–maximum range are highlighted. The median, sample size and contributing batches remain visible. Changing the selected batch or origin filter recalculates the reference.
Missing values, trace results and non-detections remain distinct; none are converted to zero. Area percentages and mass percentages are kept separate. Historical ranges are descriptive observations, not certified standards, validated anomaly probabilities, or evidence of purity or clinical suitability.
Example suppliers and reports are fictional. Imported reports remain in your browser, separate from examples. Report metadata is supplied by the importer and should be checked against the original laboratory report. Export a library backup to retain a portable copy.
The supplier catalogue uses collected source data. Product-linked reports and same-supplier botanical matches are labeled separately; neither confirms which batch will be shipped. Original documents are archived privately. Automatic extraction stays outside anomaly conclusions until reviewed, and chemical departures require at least three distinct earlier matching original reports. Unknown evidence is shown as a gap.
Open supplier directory Open GC/MS workspaceHow the Synergy Rating works
The Synergy Rating scores an essential-oil blend 0–100 as a craft / harmony / fit-to-intent + safety heuristic. It does not measure chemical synergy, absorbed dose or medical efficacy; all effect language is traditional/reported. Five quality pillars form a weighted geometric core, an evidence-qualified interaction multiplier modifies it, the Safety Gate gates it, and Data Completeness appears beside it.
Evidence-qualified Interaction multiplier
Dose-, endpoint-, route- and null-model-specific constituent observations, applied as a neutral-anchored multiplier. It stays ×1.0 when no observation qualifies; all current legacy rows are held neutral.
Aroma Harmony core
Accord structure, cohesion and note balance, computed from corpus-derived descriptor similarity.
Aroma Psychology core
A descriptor valence and arousal heuristic for odor-to-affect alignment. It does not perform airborne odor-threshold arithmetic.
Intent Alignment core
Composition-level systemic marker fit for the chosen goal, weighted by application route. It is not an absorbed-dose estimate.
Pairwise Coherence core
Constituent overlap, qualified interactions and note complementarity, weighted by oil amounts. Missing composition creates an overlap range; the score uses its conservative end so incomplete profiles cannot gain extra overlap credit.
Tenacity core
Longevity from volatility class and note anchoring, refined by measured hours and the carrier's fixative effect.
Safety Gate gate
Fail-closed, route- and population-aware safety check. A hard block caps the rating at 19 regardless of craft.
Data Completeness reported
A composition-weighted High/Medium/Low coverage band reported beside the rating. It does not enter the number.
The carrier is a first-class input
The diluent base feeds five pillars: penetration, oxidation, comedogenicity and sensitization modify the Safety Gate; delivery modifies Intent Alignment; off-notes and affect damping modify Aroma Harmony and Psychology; and the fixative effect of an occlusive base adjusts Tenacity. Safety modifiers add hazards or penalties; delivery and fixative modifiers can increase or decrease their respective scores. "No carrier" is exact identity — nothing changes.
The 99–100 product rule
Only blends rated 99–100 that also pass the engine's publish-eligibility safety gate enter the public directories. The gate is a conservative product setting, not a clinically or statistically validated threshold. Unresolved safety data blocks publication. Duplicate formulas collapse to the first discovery.
Cited data, visible gaps
Directory profiles carry source lists and visible gaps. Current runtime interaction rows are incomplete and contribute no interaction credit; many caps, thresholds and carrier fields also lack field-level verification, so a safety reviewer must approve dermal publication. Data Completeness appears beside each rating instead of hiding missing evidence inside the score.
Validation status
Formula 4.2.2 uses the report archive to test how missing constituent measurements affect chemical overlap. The possible overlap range describes missing composition; it is not a confidence interval or a prediction of benefit. The formula has deterministic regression and safety-invariant tests. The project has no independent human blend-outcome dataset yet, so those tests establish stable behavior rather than predictive accuracy. An 18-formula, 30-slot blinded study and evaluator are prepared, with no independent outcomes collected yet. Accuracy claims will require sensory panels, route-specific dose data and a locked external holdout before making an accuracy claim.