Resolve the exact UK product
Model, SKU, variant and market are resolved before comparison. GTIN is retained where the source exposes it. A retailer title is not allowed to merge similar products.
The evidence stays visible.
Every published product passes the same identity and traceability gates. The score itself changes by category because cooling, moisture removal, portable power, cooking, floor care, autonomous cleaning and audio are not interchangeable decisions.
Keep evidence classes separate
Manufacturer claims, retailer observations, controlled tests, regulatory records and owner reports remain distinct. A published specification is never relabelled as an Agent Review Data measurement. Images remain source-hosted retailer or manufacturer assets with a linked source receipt; a future affiliate feed would need the same traceability.
Gate first, score second
Exact identity, traceable specification, complete score inputs, a timestamped offer observation and resolved material conflicts are required. A numeric score cannot override a failed gate.
Rank only within a category
Each category contains 20 exact products and receives ranks 1–20. A score of 80 for a power bank has no relationship to a score of 80 for a dehumidifier.
Separate product quality from buyability
Stock is an exact-product, source-specific observation with its own checked time. Page render time is never used as evidence freshness, and generic “Add to cart” strings are not accepted as stock proof.
Deduplicate owner feedback and third-party evidence
Native ratings and counts remain source-specific. Syndicated mirrors share one pool; small or commercially hosted pools are capped and pulled toward neutral rather than presented as false internet-wide consensus. Independent tests, owner pools and Reddit discussions are labelled by evidence class and kept claim-level; they do not become a universal score.
Make commercial influence impossible
Offer redirects are affiliate-ready and attributable. Affiliate state, merchant commission and conversion performance are not inputs to evidence, eligibility, factors or rank.
Refresh atomically
Source collection validates 20 complete products per category, recomputes ranks and replaces the snapshot only after all publication gates pass. A collection failure cannot silently shrink the live catalogue.
npm run refresh:catalogue
Category contracts
Seven methods, one provenance standard
Weights below describe the current specification-led MVP. Missing controlled tests remain visible and penalised.
Portable air conditioners
Specification-led comparison of exact GB models. Published values are not treated as laboratory-equivalent measurements.
- Cooling per pound
- 28%
- Published noise
- 24%
- Claimed efficiency
- 22%
- Practical features
- 14%
- Evidence coverage
- 12%
Dehumidifiers
Specification-led comparison. Extraction claims can use different temperature and humidity conditions, so the score is directional.
- Extraction per pound
- 25%
- Published noise
- 22%
- Nameplate efficiency proxy
- 22%
- Moisture-control features
- 19%
- Evidence coverage
- 12%
Power banks
Specification-led comparison of exact UK offers. Nameplate capacity and peak output are not delivered-energy tests.
- Capacity and output per pound
- 28%
- Published maximum output
- 24%
- Nameplate capacity
- 18%
- Convenience features
- 18%
- Identity and evidence
- 12%
Air fryers
Specification-led comparison of exact UK models. Capacity, power and modes are published specifications, not cooking-lab measurements.
- Value at checked price
- 22%
- Published capacity
- 22%
- Cooking flexibility
- 22%
- Convenience features
- 22%
- Evidence coverage
- 12%
Cordless vacuums
Specification-led comparison of exact UK models. Run time is a published maximum and is not a like-for-like measured result.
- Value at checked price
- 22%
- Published run time
- 22%
- Published weight
- 22%
- Cleaning features
- 22%
- Evidence coverage
- 12%
Robot vacuums
Specification-led comparison of exact UK models. Published suction and navigation claims are not treated as a substitute for a controlled home test.
- Value at checked price
- 22%
- Published suction
- 22%
- Navigation and obstacle handling
- 22%
- Dock and mopping automation
- 22%
- Evidence coverage
- 12%
Wireless earbuds
Specification-led comparison of exact UK models. Battery life and protection ratings depend on conditions and are not independently measured here.
- Value at checked price
- 22%
- Battery life
- 22%
- Listening features
- 22%
- Protection and charging
- 22%
- Evidence coverage
- 12%
“Ranked” means comparable published evidence—not a laboratory verdict.
The catalogue is useful for shortlisting, inspecting claims and checking offers. It does not claim identical test conditions where the sources do not provide them.
Inspect the comparison tables