Use cases

Real examples of how Moresq resolves data conflicts across industries.

E-commerce — Product feed reconciliation

The problem

You aggregate product data from 8 suppliers. Each sends a different price, name, and description for the same SKU.

The solution

Moresq matches entities by SKU, resolves price conflicts with CATD, and flags "Unknown" descriptions as placeholders.

Correct price selected. Typo in supplier_b detected via fuzzy matching.

await moresq.pipeline({
  entity_id: 'SKU-12345',
  evidence: [
    { field: 'price', source: 'supplier_a', value: 29.99 },
    { field: 'price', source: 'supplier_b', value: 31.50 },
    { field: 'price', source: 'supplier_c', value: 29.99 },
    { field: 'name',  source: 'supplier_a', value: 'Wireless Headphones Pro' },
    { field: 'name',  source: 'supplier_b', value: 'Wireles Headphone Pro' },
  ]
})
// → price: 29.99 (conf 0.87), name: 'Wireless Headphones Pro' (conf 0.90)

Real estate — Multi-agency listing data

The problem

17 luxury agencies list the same properties with different prices, surfaces, and descriptions. Some in EUR, some in USD.

The solution

Moresq converts currencies, matches properties by address, resolves conflicts, and checks coherence (price/sqm ratio).

Currency converted. Coherence check flags the sqft/sqm confusion.

await moresq.pipeline({
  entity_id: 'property_monaco_01',
  evidence: [
    { field: 'price', source: 'barnes',    value: '€4,200,000' },
    { field: 'price', source: 'sothebys',  value: '$4,500,000' },
    { field: 'surface', source: 'barnes',   value: 120 },
    { field: 'surface', source: 'sothebys', value: 1290 }, // sqft, not sqm
  ],
  preset: 'property'
})
// → price: €4,200,000 (conf 0.85)
// → coherence warning: surface 1290 may be in sqft

Maritime — Yacht intelligence from 27 scrapers

The problem

27 sources report different LOA, beam, builder, and flag for the same yacht. Some say "Unknown", some use feet instead of meters.

The solution

Moresq matches by IMO/MMSI, filters "Unknown" placeholders, converts feet to meters, and resolves with CATD.

333K+ evidence reconciled. 19,756 fields updated. 8,992 placeholders filtered.

await moresq.match({
  preset: 'vessel',
  entities: [
    { source: 'yatco', fields: { name: 'M/Y Eclipse', imo: '1009613', loa: 162.5 }},
    { source: 'broker', fields: { name: 'ECLIPSE', loa: 533 }}, // 533 feet
  ]
})
// → matched by IMO, LOA conflict detected (m vs ft)

Multi-LLM — AI model scoring & routing

The problem

Your app queries 5 AI models. They return different answers. Which one is right? Which model should you use for legal questions vs math?

The solution

Moresq scores the responses, detects hallucinations, and learns reliability per model per domain over time.

After 100 queries, Moresq knows which model to trust for which domain.

const score = await moresq.ai.score({
  question: 'What is the average rent in Paris?',
  domain: 'real_estate',
  responses: [
    { model: 'claude',     extracted_value: 1200 },
    { model: 'gpt-4o',     extracted_value: 1350 },
    { model: 'perplexity', extracted_value: 1320, has_source: true },
  ]
})
// → consensus: 1291, perplexity weighted higher (has_source)

const route = await moresq.ai.route({ domain: 'legal' })
// → recommended: 'claude' (0.92 reliability for legal)

IoT — Noisy sensor data

The problem

20 temperature sensors on a factory floor. Some drift, some fail, some report in Fahrenheit.

The solution

Moresq uses Huber M-estimator (robust to outliers) and incremental O(1) mode for real-time streaming.

Sub-millisecond updates. Outlier sensors automatically downweighted.

// Incremental mode: 1µs per update
entity.addObservation('temperature', 'numeric', 'sensor_14', 22.3)
entity.addObservation('temperature', 'numeric', 'sensor_15', 22.1)
entity.addObservation('temperature', 'numeric', 'sensor_16', 85.0) // Fahrenheit!
// → consensus: 22.2°C, sensor_16 flagged as outlier

Fintech — Market data providers

The problem

4 market data providers give different stock prices, volumes, and timestamps for the same ticker.

The solution

Moresq resolves with recency-weighted consensus and cross-validates price/volume coherence.

Most recent + most agreed value wins. Stale data automatically deprioritized.

await moresq.resolve({
  field: 'price',
  strategy: 'recency',
  observations: [
    { source: 'bloomberg', value: 142.50, observed_at: '2026-03-30T10:00:01' },
    { source: 'reuters',   value: 142.48, observed_at: '2026-03-30T10:00:00' },
    { source: 'yahoo',     value: 142.50, observed_at: '2026-03-30T09:59:58' },
    { source: 'delayed',   value: 141.20, observed_at: '2026-03-30T09:45:00' },
  ]
})
// → 142.50 (conf 0.92), delayed source downweighted by recency

Try it yourself →