Per-asset agent grounding bundle — profile (composite score, ranks, returns, volatility, Sharpe), an 8×8 factor cube (t-score percentile and band), regime, and raw order-book liquidity. Single aligned snapshot.

Per-asset bundle of the latest atomic and composite signal snapshot plus N-day event lookback, grouped by asset. Includes macro regime and confidence. The full signal picture for an agent prompt in one call.

Factor t-scores for one asset across 8 classes and 8 horizons (1d-365d). Snapshot mode returns the latest anchor day; series mode returns a daily window. Standardised, z-capped, policy-aware.

Per-asset spot order-book depth timeseries — bid/ask USD at 50/100/200bp plus 1d/7d/30d spot volume. Optional futures envelope adds perp OB depth. Pre-trade depth and venue routing input.

Per-asset macro regime classification with confidence — expansion, consolidation, or deleveraging. Includes secondary regime and resolution mode. Factor-derived from the nightly 04:15 UTC pipeline.

Per-asset triggered signal feed on the latest day — every active signal with z, confidence, IC, hit rate, regime accuracy, and full health grading. Per-signal walk-forward IS/OOS block. Live firing panel with regime context.

Cross-class positioning panels per asset — futures funding stress, options 1M risk-reversal skew, order-book imbalance, and 7-day taker buy/sell ratio. Each panel: current z, percentile, 30-day sparkline, one-line narrative. Response carries a _next block with active-signals / thesis / regime-context / top-setups / peer-cluster so an agent can go from positioning read to per-asset conviction, historical edge, or basket construction without leaving the flow.

Per-asset metric health grades for one symbol — TS vs. CS grade divergence, per-metric ts_ic, ts_hit_rate. Answers "which metrics actually work on THIS asset", not just the universe average. Companion to /asset/top-predictors which ranks; this endpoint diagnoses.

Nearest k peers for one asset in CRYPTYX's 8-class factor space at the requested horizon. Returns target asset metrics plus peers ranked by Euclidean distance with divergence stats. Hedge selection and basket construction.

Per-asset regime state across all 8 horizons plus divergence flags vs. macro. Enhanced view over /asset-regimes: adds asset-vs-market comparison and per-horizon divergence indicators — gates whether the asset is in-step or out-of-step with the broader market at each timescale.

Per-asset macro regime context — current regime id, confidence, days-in-regime streak, conditional forward-return stats at 1d/7d/14d/30d, and top historical regime transitions. Grounds sizing in what historically happens next.

Per-asset triggered signal ledger over 7-365 days, A/B grades only. Per event: signal, direction (long/short/neutral), z, confidence, horizon, grade, close price, and realised 7-day forward return.

Claude-narrated per-asset conviction thesis (3-5 sentences) plus an evidence pack — composite rank vs universe, returns and volatility, factor scores per class, top active signals, and top per-asset predictors. 4-hour cache.

Per-day composite conviction trajectory for one asset over 7-90 days. Score timeline always complete; optional narrative one-liner overlay from the daily thesis job. Track conviction evolution before a major move.

Per-asset predictor explorer — top 5 atomic signals and top 5 metrics ranked by absolute IC times hit rate. A/B grades only, anti-predictive filtered, family-deduped. Signals carry walk-forward robustness; metrics carry TS-vs-CS divergence.

Historical signal edge for one asset and horizon — up to 6 A/B-graded atomic signals that actually worked on this asset, grouped into reversal/continuation/breakout lanes. Per setup: hit rate, mean forward return, walk-forward IS/OOS.

Per-asset snapshot or daily series — composite score, rank, recent returns, realised volatility, Sharpe ratios, plus macro regime with confidence. The 200+ asset master ranking feed for portfolio construction and rotation.

Aggregated-exchange daily perpetual funding rate per asset, last N days. Raw feed — no annualisation, no z-score. Cheap historical input for carry-trade backtests and crowding diagnostics.

Natural-language entry point to the full CRYPTYX stack. POST a query string; an LLM classifies into 60+ intents (conviction_ranking, market_brief, factor_breakdown, signal_explain, regime_check, top_setups, etc.) and routes to the matching pre-computed analytics over 200+ assets, 152 signals, 440+ metrics. Response payload is intent-shaped — a top-3 conviction query returns a compact ranking; a market_brief returns a full universe snapshot. Wire an LLM agent into CRYPTYX without plumbing 30+ endpoints.

Latest-day Deribit IV term structure for one asset — per-tenor ATM IV, 25-delta risk reversal, butterfly, plus current put/call OI ratio. BTC and ETH only. Feeds the OPT factor class; use for skew reads and vol-term structuring.

Daily OHLCV bars for one asset over 1-1825 days — open, high, low, close, USD volume. UTC-1 anchor day. The grounded spot tape behind every CRYPTYX factor score and signal.

Universe-wide factor breadth bucketed by class and horizon — positive/negative/neutral counts, average t-score, expand-share percent, dominant regime. Time risk-on/off rotations across 200+ assets.

Latest-day breakdown of composite signal trigger stats: per-composite triggered_asset_count, breadth %, top-triggering assets. Snapshot of where the 9 IC-weighted composites are actually firing across 200+ assets right now. Complements /signals/composite/breadth (universe-wide) with per-composite drill-down.

Daily "where to look" leaderboard — top-N and bottom-N assets by composite score with returns, funding/OI z-scores, signal counts, optional thematic scores, and macro regime gating. Stablecoins filtered. Response carries a _next block with thesis / active-signals / asymmetry / regime-context / trade-ideas so an agent can go from "which asset" to "what does the full CRYPTYX stack say about it" in a single hop.

Universe-aggregate cross-factor divergence alerts — Ignition (TR+VOL extreme), Capitulation (FLOW+TR collapse), and Distribution (outflows with trend). Severity-ranked by horizon for top-down scanning.

Factor term-structure grid — 7 classes by 8 horizons of market-average t-score with breadth counts and 1-day momentum delta. Per-asset mode adds delta-vs-market per cell. Read rotation in one matrix.

BTC price + macro regime history + high-conviction signal density timeseries in a single call (1-1900 days). Bundled dashboard hero for orienting on "where are we in the cycle" — configurable horizon (1d-365d) and layer (core/signals). One call replaces 3 separate history queries.

Multi-horizon macro regime synthesis from 1d through 365d — regime label, confidence, cross-class state, breadth, horizon alignment. Read full risk-on/off posture in one call. Slim mode drops blobs for timelines.

Historical performance of a specific (horizon, regime_label) cell — n_prints, mean 30d forward return, hit rate, and last 5 instances. Answers "when this regime showed up before, what happened next?" — analytical primitive for regime-conditional strategy backtesting.

Market vs. asset regime cascade divergence for one symbol — score delta, per-horizon breakdown, severity, and one-line interpretation. Novel cross-asset positioning read: "where is this asset out-of-step with the macro?" Companion to /market-pulse/divergences (which scans the universe); this endpoint drills into one asset.

Tactical / strategic / secular regime decomposition + Supertrend score across a 1-1900 day lookback. Shows regime alignment across timescales — is the market in agreement or diverging? Complements /market-pulse/regime by decomposing multi-horizon consistency into named layers.

Lightweight 90-day z-score timeseries for one (metric_id, asset). Returns asof_day + z + percentile per day. Cheap primitive for chart tooltips, sparklines, or rapid metric due-diligence before committing to a full /metrics/slicer backtest at $0.05.

Single-metric z-score backtest over 5+ years of history per asset. Triggers, trigger rate, and per-horizon forward-return and max-drawdown stats across 8 horizons (1d-365d). Pre-composite single-factor edge testing.

Multi-factor z-score confluence backtest for one asset (2-4 conditions). Intersects triggered days, returns per-metric series, composite triggers, and per-horizon forward-return/drawdown stats. Validate confluence stacks before live deployment.

Cross-sectional z-score scanner across 200+ assets and 440+ metrics. Pick a metric and threshold; get ranked triggering assets with z, percentile, returns, composite score, and per-metric health grading. Live anomaly hunting.

Today's firing signals across 200+ assets — atomic and composite arrays with confidence, macro regime, and event-deltas vs the prior day. Anti-predictive signals filtered. Response also carries a _next block pointing at explain / top / market-pulse/conviction / active-signals / trade-ideas so you can drill in without leaving the flow.

Backtest any signal over up to 365 days. Returns trigger stats, per-day rows, and a 4-horizon performance block (sample size, mean/median return, hit rate, bootstrap CI, statistical significance, reliability grade). Walk-forward OOS read inlined.

The entry point for agents landing on CRYPTYX. Returns the full signal registry (143 atomic + 9 composite) with class, horizon, geometry, parameters, 30-day trigger stats, three-axis health grades, walk-forward IS/OOS — plus a meta.service block (universe + walk-forward + macro-regime + OKX/Kraken/Coinbase venues), a meta.by_class strategy-family breakdown, a meta.next recommended 5-step flow, and per-signal _actions with direct URLs to explain / backtest / fork / simulate. One call answers "what is CRYPTYX, what can I buy, and what should I call next".

Class-level attribution waterfall for one asset and horizon — master score, per-class contributions ordered by impact, per-signal triggers with walk-forward IS/OOS, and per-signal health grades.

Walk-forward backtest of a custom 8-class composite weight vector at 7d/14d/30d. Runs custom, IC-weighted default, and equal-weight blends in parallel with full performance breakdowns. Benchmark your blend against the production moat.

Universe-wide composite breadth at one horizon — bullish/bearish/neutral counts and percentages across 200+ assets, sliced by regime, with full per-asset stance list. Is this rally broad or concentrated?

Re-rank 200+ assets under your own 8-class composite weights at 7d/14d/30d. Returns full ranking with class contributions, default-weight comparison, and top movers vs default. Stress-test your composite recipe against the production moat.

Latest-day composite snapshot — master composite scores per (asset, horizon) with ranks, per-class composites with trigger counts, macro regime context, and intraday micro-regime state. Multi-horizon factor heatmap input.

N-day rolling delta on composite scores (master + per-class) with direction label (building / fading / steady). Detects conviction acceleration or decay — "is this composite gaining or losing energy?" Complements /signals/composite/heatmap (levels) with change-detection.

Re-run the database-side signal evaluator for one day. Dryrun returns predicted trigger counts; live run executes and returns inserted row counts. Pipeline smoke test or recover a missed eval.

Factor t-score breakdown for one (asset, day) — per-class scores across the 8 factor classes plus the weighted composite. Inspect which factor classes drove the call.

Fork a signal into an inactive parameter variant — production firing is untouched. Experimentation primitive: fork, dryrun-backtest, then activate. Returns the new version id and parameters.

Signal activity leaderboard over 7-365 days — per-signal trigger frequency, average confidence, assets triggered, and a per-horizon walk-forward IC/reliability block. Frequency and robustness in one pass.

Recently triggered atomic signal events (up to 50, last 1-7 days). Per event: asset, signal_id, confidence, asof_day. Change-detection digest for monitoring. Response carries a _next block with explain / catalog / top / thesis / trade-ideas so an agent can pivot from event stream to per-event drill-down or asset thesis in one hop.

Pre-flight threshold simulator — count how many assets would trigger at a proposed cutoff for one anchor day, with current vs proposed threshold and estimated trigger rate. Use before forking parameters.