Every product in the Vyvata catalog gets scored on a 0-100 scale. The score determines which tier the product appears in (Rejected, Standard, Verified, Elite). It also determines whether we recommend the product in personalized protocols, how the product ranks in category-level searches, and whether it appears in curated collections.
The score is not a vibe. It is the output of a deterministic rubric across six weighted dimensions, with published auto-fail rules that immediately disqualify products with specific integrity problems. The methodology is the same for every product in every category. There is no editorial thumb on the scale.
This is the deep dive into how the whole thing works. It covers the six dimensions, the auto-fail rules and why they exist, the tier band cutoffs, the difference between how supplements and non-supplements get scored, and the transparency-first policy that leaves Rejected products listed with an explanation rather than hidden. If you want to know why any specific product got its score, this is the document that lets you audit it.
Why we score at all
The health-tech and wellness market has a fundamental transparency problem. Products are sold with a combination of marketing language, testimonial evidence, and regulatory-flavored phrases that mostly aren't verifiable. A consumer trying to decide between two red light therapy devices, two blood pressure monitors, or two magnesium supplements has almost no way to compare them without spending days on research per product.
The Vyvata Score is our attempt to compress that research into a single number. Not because a single number is sufficient — it isn't — but because it forces the underlying evaluation to be reproducible. If two evaluators run the same product through the rubric, they get similar scores. If a manufacturer changes their documentation and we re-score, we can explain exactly what changed. The score is the audit trail.
The alternative — editorial reviews that reflect the reviewer's judgment without an underlying rubric — leads to inconsistency across products, category, and time. It also creates conflicts of interest that are hard to detect. We wanted a system where the incentives were legible, and the trade-off is that the system is more mechanical than a Wirecutter-style curated review. That is a feature, not a bug.
The six dimensions in order of importance
Every product is scored on six dimensions. The dimensions have different weights based on how much each contributes to whether a product does what it claims. The weights are fixed — the same for every category.
Evidence
The largest dimension. Evaluates whether the product's core claims are supported by peer-reviewed research at doses and configurations that match the actual product. For medical devices, this includes FDA clearance and clinical validation. For supplements, it includes ingredient dose evidence against the specific formulation. For wellness devices, it includes modality-level evidence (red light photobiomodulation, sauna heat therapy, cold exposure) combined with product-level specification adequacy.
Products earn evidence credit for citing specific studies, publishing FDA clearance numbers, and matching their formulation to trial-supported doses. They lose credit for vague claims ("clinically shown," "medical grade"), for citing ingredients without matching the effective dose, and for making disease claims without the regulatory backing to support them.
The weight reflects the principle that a product with weak evidence is not worth much regardless of how nicely it's built. A beautifully designed device with no efficacy is still an ineffective device.
Formulation
Evaluates the specific composition of the product. For supplements: per-ingredient dose disclosure, absence of proprietary blends that hide dosing, form of the compound (magnesium glycinate vs. oxide matters), and appropriate excipients. For devices: hardware specifications, sensor accuracy, precise dose delivery (irradiance in mW/cm² for LED devices, pressure in cmH2O for CPAP), and build tolerances.
A magnesium supplement listing 400mg magnesium glycinate scores higher than one listing "200mg proprietary blend of magnesium." A red light panel publishing 660nm at 100 mW/cm² at 6 inches scores higher than one saying "red and near-infrared wavelengths." The dimension rewards specificity because specificity is what makes evaluation possible.
Manufacturing
Evaluates the manufacturing quality and supply chain integrity. For supplements: cGMP compliance, third-party testing (USP, NSF, Informed Sport), and Certificate of Analysis availability per lot. For devices: manufacturer identity, ISO 13485 quality management for medical devices, UL/ETL/CE electrical safety certifications for powered devices, and country-of-origin transparency.
This dimension is what separates real brands from Amazon-facing resellers. A named manufacturer with an ISO 13485 facility and published quality controls scores well. A dropship storefront with no manufacturer identity scores poorly regardless of the underlying hardware quality.
Safety
Evaluates safety documentation and adverse event tracking. For medical devices: MAUDE (Manufacturer and User Facility Device Experience) database records, FDA recall history, and post-market surveillance quality. For supplements: contraindication disclosure, drug interaction warnings, and dose-limit clarity. For all products: emergency safety features (auto-shutoff, thermal protection), user warnings, and manufacturer response to reported issues.
Products with clean MAUDE records and prominent safety documentation earn credit. Products with recalls, injury reports, or missing contraindication information lose credit.
Transparency
Evaluates the accessibility of product documentation. This is the dimension that most distinguishes brands that are proud of what they sell from brands that hide behind marketing. Specific criteria include: privacy policy, terms of service, return and warranty policy accessibility, corporate identity, contact information, and machine-readable documentation.
The transparency score is what our automated scoring engine can most reliably measure — is the FDA clearance in HTML on the page, or buried in a PDF? Are the ingredient doses in structured data, or in an image? Are the third-party COAs linked from the product page, or only available on request? If we can't verify it programmatically, we treat it as absent.
Sustainability
Evaluates the environmental and social footprint. Recyclable packaging, ethical sourcing, energy efficiency for powered devices, replaceable components for long product life, and labor practices where disclosed. The smallest dimension because it affects consumer outcomes less directly than the others, but non-zero because it matters over time and at scale.
Products with detailed sustainability documentation earn credit. Products with no disclosure default to average scores rather than punitive lows — the absence of documentation isn't necessarily evidence of bad practice, but the presence is evidence of good.
The four auto-fail rules
Some product problems are severe enough that they should immediately disqualify a product regardless of how well it might score on the individual dimensions. We use four specific auto-fail rules to catch these.
1. Unverified medical claim
Triggers when a product claims a medical function that requires FDA clearance (blood pressure monitoring, blood glucose measurement, ECG, AFib detection, disease treatment) without publishing a specific 510(k) clearance number or equivalent regulatory documentation.
This is the most common auto-fail in the smartwatch and ring categories. Products claiming ECG or blood pressure monitoring without clearance immediately fail regardless of other dimensions. The F300 smartwatch and dozens like it fall into this category.
2. Misleading naming
Triggers when a product's name co-opts a regulated device category or brand it does not legally represent. "Mini CPAP" for a device that generates no positive airway pressure. "Xiaomi Smart Ring" for a product that is not manufactured by Xiaomi. "Calm sound machine" using the trademark of an unrelated meditation brand.
This auto-fail exists specifically because the naming trick can cause direct patient harm — a person with sleep apnea trusting a "Mini CPAP" as a substitute for real therapy is at genuine cardiovascular risk. The Mini CPAP nasal device in the catalog scored 15 for exactly this reason.
3. No brand provenance
Triggers when a product cannot be traced to a specific manufacturer with meaningful corporate identity. Generic dropship listings, Amazon-facing resellers using invented brand names, and products where the same reference SKU appears under multiple storefront names all fall into this category.
The Vyvata catalog has dozens of these rejections. The underlying hardware may be functional. Without a manufacturer accountable for quality control, safety, and post-market response, we cannot certify the product to any tier.
4. Out of scope for biohacking
Triggers when a product does not fit any category in Vyvata's taxonomy. Cosmetic beauty devices marketed as biohacking. Pet products. Fire suppression accessories. Metaphysical intention kits. These are auto-rejected because there is no honest way to score them against our rubric — the rubric wasn't designed for them.
The tier bands
Scores map to four public tiers.
-
Rejected: 0-49. The product either triggered an auto-fail or scored low enough on the weighted dimensions that we cannot honestly recommend it. Rejected products stay listed in the catalog with an honest explanation of why they scored low.
-
Standard: 50-69. The product is functional but has meaningful gaps in evidence, formulation, manufacturing, safety, or transparency. It may be worth buying for specific use cases, with eyes open about the tradeoffs.
-
Verified: 70-84. The product meets our documentation and evidence thresholds. Strong build, real regulatory backing where applicable, transparent manufacturing. Worth buying without heavy caveats.
-
Elite: 85-100. The product is exceptional across dimensions. Very few products reach this tier. As of writing, the Dexcom Stelo at 91 is the highest-scoring product in the catalog.
The band boundaries are round numbers by design. Not because 69 versus 70 is a meaningful qualitative difference, but because the boundaries need to be defensible and predictable. A product one point below the Verified threshold is not meaningfully worse than one at the threshold — but the threshold has to sit somewhere.
Why some products score 20 and others score 90
The score range is not a smooth distribution. Products cluster around specific bands based on how they engage with the rubric.
Products at the 15-30 range are typically auto-fail rejections. The Mini CPAP at 15 triggered misleading_naming. The F300 smartwatch at 20 triggered unverified_medical_claim. These products may have functional hardware but they have integrity problems that make them ineligible for a positive score regardless.
Products at the 30-45 range are typically weak across multiple dimensions. Vague medical claims, missing manufacturer identity, no third-party testing, no safety documentation. The score reflects that we can't verify almost anything about them.
Products at the 50-70 range are typically legitimate but under-documented. Real brands with functional products but missing specific certifications, weaker evidence bases, or incomplete transparency. The Nushape Mini Red Light Therapy Device at score 66 (Standard) is a representative case — a fine device from a real brand, but with less regulatory documentation than the Lumara VISO at 82.
Products at the 70-85 range are the Verified backbone. Strong on most dimensions, some minor gaps, worth buying. The Withings ScanWatch 2 at 85, the Theragun Mini Plus at 78, the Lumara VISO at 82. These are the products we point beginners toward.
Products at 85+ are rare. The Dexcom Stelo at 91 earns its Elite tier through the combination of FDA clearance, extensive clinical validation, transparent MARD accuracy data, and manufacturing pedigree.
The transparency-first policy: why Rejected products stay listed
Most product review sites hide the products they don't recommend. A Wirecutter review lists the top pick and the runners-up; the fifty products they evaluated and rejected are invisible.
Vyvata does the opposite. Every product we score, including the Rejected ones, stays listed in the catalog with the score visible and the auto-fail explanation public. If you search for "CPAP" on the site, you see the Mini CPAP nasal device with a score of 15 and the explanation "passive plastic nasal dilator marketed as a Mini CPAP" right next to it.
This policy is deliberate. The reasoning is:
-
Consumers still search for these products. If someone is going to buy a Mini CPAP, we would rather they encounter our scoring first than end up on an Amazon listing with no critical evaluation.
-
The receipts matter. Publishing the specific reason for rejection teaches readers how the rubric works. "Rejected on misleading naming" is more educational than the product simply being absent.
-
Editorial pressure is legible. A brand that wants to be re-scored has to fix the specific documentation problem we cited. This is a public feedback loop.
-
Trust compounds over time. A catalog that lists both what earned Verified and what earned Rejected is more trustworthy than one that only shows the picks.
Supplements versus non-supplements: the scoring difference
The rubric is the same, but the source data differs meaningfully between supplement and non-supplement products.
Supplement scoring is heuristic and deterministic. We evaluate the product listing against the six dimensions using structured criteria: per-ingredient dose disclosure (checked programmatically), third-party testing certification (checked against known cert databases), form of the compound (parsed from the ingredient list), and manufacturer identity (checked against the FDA facility registration database). The scoring runs deterministically — the same product listing produces the same score every time.
Non-supplement scoring is LLM-driven with API integration. For medical devices, we query three large national databases for validation studies matched to the product's specific configuration. For consumer wearables, we cross-reference against known databases and registries to identify reference designs. The LLM evaluates the product page against the rubric using this enriched data.
The LLM approach is necessary because non-supplement products vary too much for pure heuristic scoring — a red light panel and a sauna and a smartwatch have almost nothing in common at the specification level, but the same six dimensions still apply. The LLM interprets each dimension in context.
Both approaches are reproducible. Both are audited against a manual reference set periodically. Both produce scores that fit the same tier bands.
What the score does not measure
The rubric is not comprehensive. Specific things it deliberately does not weigh:
-
Price. A cheap product and an expensive product can both earn Verified if they meet the criteria. Value-per-dollar is a downstream question the user makes based on their budget.
-
Aesthetics. How the product looks does not factor. A functional ugly device outscores a beautiful ineffective one.
-
Brand popularity. Trendy new brands don't get lifted by hype. Established brands don't get penalized for being unfashionable.
-
Marketing spend. A brand's advertising presence has no bearing on the score. This is deliberate.
-
Editorial preference. The rubric is designed to make individual preference irrelevant. Two evaluators running the rubric should reach similar scores.
What happens when a brand challenges a score
Brands regularly contact us to challenge Rejected scores. We take these seriously. The process is:
-
The brand provides the specific documentation they believe was missed — FDA clearance number, third-party test COA, manufacturer verification.
-
We re-run the scoring with the added documentation.
-
If the new documentation changes the underlying dimension scores or removes an auto-fail, the score updates.
-
The explanation in the receipts file updates.
What we do not accept: arguments that the auto-fail rules should not apply, requests to hide the Rejected listing, or offers to compensate for score updates. The rubric is the rubric. The way to score higher is to publish the documentation that makes the rubric grant credit.
The bigger picture
The Vyvata Score is not perfect. No rubric is. Some products fall in awkward gaps between the dimensions. Some auto-fails may occasionally catch a product that would be reasonable in context. Some Verified products may turn out to have quality issues we didn't catch. We update the methodology as we find problems, and we publish the updates.
What the rubric does well: it is transparent, reproducible, and free from most of the conflicts of interest that shape traditional product reviews. It generates scores that survive audit. It leaves the Rejected products listed with honest explanations. And it lets a consumer with 60 seconds get a defensible answer to "is this product worth buying," backed by criteria they can inspect if they want to.
The methodology page has the full technical specification. The Search hub lets you browse products at each tier. And the Find Your Protocol quiz uses the same scoring under the hood to generate personalized shortlists. Everything is deterministic. Everything is auditable. If you find a score you disagree with, tell us why — the methodology is designed to be updated when the criticism is specific enough to act on. That is the deal we're making with users.
