The discipline you'd apply to one card, applied to every card, every hour.

Vision extracts what it can read off the card: subject, set, number, parallel, grade, cert. Then the cascade fires every authoritative source that applies in parallel. PSA cert number? Resolved exactly. Magic card? Scryfall returns the canonical printing. Pokémon? pokemontcg.io. Marvel issue? ComicVine. Each source returns a confidence score so the UI can badge how it knows.

When trigram has nothing but the cascade is confident, Cardpulse materializes the catalog row on the spot — authoritative image, set code, rarity, language — and routes you straight to the "Add to inventory" flow. The contribute form only shows up for cards no API has ever heard of.

A 1/1 only sells once a lifetime. A /5 might sit untraded for a year. The standard comp ladder refuses on these — and most pricing tools either go silent or fabricate a number with vague methodology. We do neither.

Cardpulse's borrowing graph pulls comps from four explicit edges: prior-year siblings of the same insert, same-year cross-player comps adjusted by a learned player-tier multiplier, the player's broader 1/1 ceiling, and a pop-class anchor multiplied by a fitted scarcity ratio. Every edge runs through a tricube kernel weighted by recency, distance, and tick suspicion. The result lands at a posterior shrunk toward a product-line prior, with the band inflated by an explicit borrowing distance — a borrowed estimate is structurally wider than a direct-comp one and never claims otherwise.

The panel shows you exactly which edges fired, their weight share, and the comps each one used. Click through to inspect every comp the borrow leaned on. No black box.

You know the moves. The unverified buyer who paid 50% over comp because they're pumping a forum thread. The shill bid that closed on a buddy account. The mis-graded card that slipped through and anchored a Sunday night BIN you were never supposed to see.

Cardpulse runs a modified z-score across every variant's tick window — the same statistic forensic accountants use for messy real-world data, not the textbook 1.5×IQR fence that breaks on small n. Each tick gets scored. Anything past |3.5| is flagged and lifted out of the median, never silently dropped — the outlier still appears in the comp ladder so you can decide.

A clean comp ladder tells you what a card was worth yesterday. Cardpulse also reads what the market is about to feel about it today — the four-touchdown MNF, the pop-report drop, the rotation rotation that just put a Pokémon archetype back in tier-one, the MCU release date that resurfaces a Silver Age key.

The agent fans out across forums and news (Reddit, ESPN, X), pulls posts about the card's player or subject, scores each one on a -1 to +1 axis with a confidence weight, and aggregates them into the forecast — but not before they've been audited (next section).

The rules pipeline has already produced a Buy / Hold / Sell. The reviewer sees the same evidence — the actual signal text, the card identity, the projection magnitude — and gets one permission: downgrade Buy/Sell to Hold when the evidence reads thin. It can never upgrade. Caution-only.

What this catches: a single ESPN injury squib for a known load-managed star. A trade rumor with no named counterparty. Modern news on a vintage HOF rookie that trades on legacy, not weekly availability. The reviewer's rationale replaces the formulaic one when it overrides — so you see "One missed practice doesn't move a PSA 10 base," not "30d trend projects -8%."

Where most pricing tools stop — at a single number — Cardpulse keeps going. Every variant gets a 30-day and 90-day forecast as a true 80% prediction band, not a fudged ±8%. Inside the band: a buy / hold / sell conviction and a one-sentence agent rationale.

Inside the rationale: the layered drivers. A 30-day regression baseline. A peer-borrow from comparable variants when the card itself is thin. A sentiment drift from the agent's signal aggregate. A grade-tier multiplier (top tiers amplify sentiment). A rookie premium (RCs trade on narrative arc). A rarity premium for short-print and 1-of-1 variants. A calendar bump when a postseason game or set release sits inside the window. Every layer carries its own contribution and confidence. No magic — just disciplined attribution.