TypeScript 61.9%
HTML 37.2%
SQL 0.7%
-
Fri, Sep 11, 2026 3
-
Auction intelligence (web): /auctions rebuilt (open / ending soon / below RIV / results / houses, filters, all-in estimates, bid vs RIV), auction-house pages with buyer-premium schedules and stats, asset Auctions tab, all-in price + fee-basis pills in sales tables, methodology section 16
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Terminal features: screener, methodology center, market depth & time-to-sell, sale verification labels, homepage hierarchy
…
- /screener: filter builder (category, grader/grade, RIV, confidence, liquidity, rarity, 1M/1Y change, sales, listings, spread, drawdown), 6 presets, sortable dense table, CSV/JSON export (5 000 rows, as_of + attribution), plausibility gates on every column - /methodology: 15-section center wired to the valuation constants; ScoreExplainer on asset pages (liquidity, rarity, confidence, ask vs RIV) with the asset's real inputs - market depth (±5/10/20 % bands, best ask vs RIV), days on market and time-to-sale by ask band from observed listing lifecycles (category-level fallback, labelled), fair buy / fast-typical-patient sell ladder; GET /v1/assets/:id/depth - sales verification badge (verified / likely / unverified / excluded, heuristic, documented) - ticker shows only published indices; 'in development' indices moved to a collapsed line; homepage reordered (RARE → snapshot → trending → records → radar → opportunities → auctions → latest → categories); trending rail quality gates; /indices and /portfolio redirects - worker: ATH/ATL/drawdown scoped to the representative variant - catawiki: staleness judged at observation time (fixes a date-dependent test) - docs: stale PENDING-SCHEMA files removed; METHODOLOGY.md + API.md updated Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Valuation safety (phase 1): ask-vs-RIV anomaly gate, variant-aware discounts, grade parsing fixes
…
- scores: assessAsk()/dealScore()/plausibility gates — an ask is compared only with a transaction-based RIV of the SAME variant (>=5 sales, confidence >=0.5, match >=0.7); asks <10% or >10x RIV are flagged riv_anomaly, never surfaced as deals - valuation worker: discounts reset each pass, representative variant prefers the default (raw/base) variant when it qualifies, changes measured on the representative variant's own series, no cross-variant fallback for value_opportunity, numeric(8,6) overflow guard - taxonomy: parse PSA/DNA n, (8.5), Authentic-only slabs; a grader without a readable grade gets its own 'grader|unknown' variant instead of polluting raw - resolver/writers: refineGrade() re-parses titles for sales, listings, observations, lots - web: VsRiv primitive (Data/identity anomaly label), opportunity/gainers/losers/listings sort + filter directions fixed (negative = below RIV), scanner verdict via assessAsk with FX - indices: RARE publishes only when its 25-constituent minimum is met - security: atomic ai_quotas upsert with unique key, scanner choose ownership check, anonymous research thread authz - db: migration 0006 (hot-path indexes, ai_quotas unique); ri regrade repair command Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
-
Mon, Sep 7, 2026 2
-
Frontend: design system, mobile shell + tab bar, search sheet, home tiles, touch charts, asset page, compare, account polish (agents P, Q)
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Public market web: home, markets, indices, explore, asset pages, search, SEO (agent E)
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-