@cryptyx/mcp-server is an institutional-grade digital asset intelligence platform offering 21 tools for quantitative research, signal management, and market analysis across ~200 crypto assets.
Factor Discovery & Research
Retrieve top-performing metrics ranked by information coefficient (IC) with conviction grades (get_featured_metrics)
Backtest single metrics via z-score analysis across 8 forward-return horizons (1d–365d)
Define 2–4 metric conditions and backtest multi-factor intersections simultaneously
Scan a metric across the entire asset universe for z-score extremes on the latest day
Retrieve t-scores for an asset across 8 factor classes (CORR, EFF, FLOW, FUT, OB, OPT, TR, VOL)
Signal Engine
Browse the full signal catalog with parameters and 30-day trigger statistics
Get today's active signal triggers with confidence scores and composite rollups
Explain why a signal fired (or didn't) for a specific asset on a given date
Backtest signals over historical date ranges to evaluate trigger rates and performance
Fork signals to create inactive parameter variants for experimentation without affecting live signals
Simulate threshold changes to estimate trigger rate impact without making actual changes
Market Intelligence
Get composite scores, returns, and rankings for the asset universe (latest or time series)
View factor breadth across the universe showing positive/negative/neutral asset counts per factor class
Get a full AI-optimized snapshot including breadth, top/bottom rankings, signal summary, and pipeline status
Classify asset market regimes (trending, mean-reverting, volatile) with confidence scores
Retrieve daily OHLCV price history and live spot prices (refreshed every 15 minutes)
Search and list all tracked assets with universe tags
Execution Context
Analyze order book depth (bid/ask USD at 50/100/200 bps from mid) for spot and futures markets to size real-world trades
CRYPTYX Challenge
View competition rounds with rules, asset universes, and entry counts
Access live leaderboards with Sharpe ratio, total return, max drawdown, and composite scores for benchmarking AI trading agents
Provides crypto trading intelligence and conviction signals that can be used alongside OKX's execution capabilities, enabling AI agents to make informed trading decisions while OKX handles order execution.
🔔 Deprecation notice (2026-07)
This package is entering deprecation. The flagship CRYPTYX MCP integration is now the OAuth 2.1 Remote MCP surface at
https://cryptyx.ai/api/mcp— no config file editing, no API key management, revocable per-app, enterprise/SOC 2 ready.Migrate at cryptyx.ai/mcp. Paste one URL into Claude.ai, Cursor, or any MCP-compatible client and go.
This npm package remains supported through 2026-12-31 for clients that don't yet support OAuth MCP (older Cursor releases, some enterprise agent platforms). Your existing
cx_*API key continues to work throughout — no forced migration timeline for existing users.New users: use the OAuth path from day one. Existing users: migrate on your schedule. See migration guide (or just re-add via
cryptyx.ai/mcp— takes 30 seconds).
CRYPTYX — the intelligence layer for digital assets.
Institutional-grade digital asset intelligence delivered via the Model Context Protocol and a human-facing CLI. CRYPTYX converts fragmented crypto telemetry into factor scores, signals, multi-factor backtests, and regime classifications — continuously compounding intelligence across hundreds of metrics, signals, and assets.
Not a data proxy. A quant research platform. 30 tools across 670+ metrics, 8 factor classes, 160 signals, 17 IC-weighted composites, ~200 tracked assets, and a daily-updating signal registry — now including institutional trigger evaluation (Sortino, MDD, profit factor, walk-forward IS/OOS) as of v0.8.0. Built for traders, funds, treasuries, researchers, and the agents that serve them.
Execution is complementary — and venue-neutral. Use CRYPTYX alongside any exchange execution rail. Reference handshakes shipped for Coinbase, OKX, Kraken, Binance, and Hyperliquid — each ~150 lines, one three-step pattern: pay CRYPTYX → apply institutional gate → execute on the venue. Fork any of them for your target venue.
Related MCP server: fathom-fyi
npx @cryptyx/mcp-servernpx cryptyx --helpOr install globally:
npm install -g @cryptyx/mcp-server
cryptyx snapshotBoth entry points ship from the same package — one install, two interfaces.
# Save your API key (persists to ~/.config/cryptyx/config.json)
cryptyx config set-key cx_your_api_key
# Full state snapshot — factor breadth, rankings, signal summary
cryptyx snapshot
# Market data for all assets
cryptyx market
# Factor scores for a single asset
cryptyx factors BTC
# Today's signal firings
cryptyx signals
# Top 10 signals by information coefficient
cryptyx signals top
# Backtest a signal over the last year
cryptyx signals backtest TR_WIN_RATE_60D_THR --from 2025-05-28 --to 2026-05-28
# Scan a metric across the universe for z-score extremes
cryptyx metrics scan VOL_RV_7D --operator gt --threshold 2
# Raw JSON output (pipe to jq, scripts, cron)
cryptyx market --json | jq '.[] | select(.composite_rank <= 10)'Every command supports --json for raw output and --api-key / --api-url overrides.
Command | Description |
| Full state snapshot: factor breadth, rankings, signal summary |
| Asset universe with composite scores and returns |
| Today's active signal triggers |
| Top 10 signals by 7d IC |
| All signals with parameters and 30d stats |
| Why a signal fired (or didn't) |
| Signal backtest over date range |
| Fork a signal with new parameters |
| Estimate trigger rate for threshold change |
| Factor t-scores across 8 classes |
| Regime classification with confidence |
| Market-wide macro regime classification across all horizons |
| Cross-factor divergence alerts (distribution, capitulation, ignition) |
| Factor breadth across the universe |
| Live spot prices (15m refresh) |
| Daily OHLCV candles |
| Full tracked universe with tags |
| Order book depth at 50/100/200bp |
| Top metrics by IC |
| Single-metric z-score backtest |
| Universe-wide z-score scan |
| Multi-factor intersection backtest |
| Competition leaderboard |
| All competition rounds |
| Save API key |
| Show current config |
| Clear saved config |
{
"mcpServers": {
"cryptyx": {
"command": "npx",
"args": ["@cryptyx/mcp-server"],
"env": {
"CRYPTYX_API_KEY": "your-api-key"
}
}
}
}claude mcp add cryptyx -- npx @cryptyx/mcp-server
export CRYPTYX_API_KEY=cx_your_keyEndpoint: https://cryptyx.ai/api/mcp
Transport: Streamable HTTP (JSON-RPC 2.0 over POST)
Auth: Bearer token (cx_* API key)Variable | Required | Default | Description |
| Yes | — | API key from cryptyx.ai |
| No |
| Override for self-hosted deployments |
CRYPTYX is designed for a specific agentic workflow. Most tools map to a step in this loop:
DISCOVER → DEFINE → VALIDATE → SCAN → STORE → EXECUTEDISCOVER — get_featured_metrics surfaces the current top-performing metrics by information coefficient (IC). Start here.
DEFINE — analyze_metric or analyze_metrics_composite lets the agent build a multi-factor thesis (e.g. "trend momentum z > 1.5 AND funding stress z > 2.0").
VALIDATE — The same tools return forward returns at 8 horizons (1d to 365d). The agent sees whether the thesis has edge, not just vibes.
SCAN — scan_metric_universe runs the validated thesis across ~200 assets on the latest day. Which assets match the conditions right now?
STORE — fork_signal persists the thesis as a new inactive signal variant. The daily pipeline will track it forever.
EXECUTE — CRYPTYX doesn't execute. Hand off to OKX, Kraken, or whatever execution layer your agent uses.
The core value of CRYPTYX. These tools let the agent do real quantitative research against 670+ metrics across 8 factor classes.
Tool | What it does |
| Top-performing metrics by information coefficient. Returns the 8 highest-conviction metrics with A/B grades. Best starting point. |
| Per-asset predictive power — metrics ranked by time-series IC for a specific asset. Surfaces hidden alpha (CS Grade F but TS Grade A). |
| Single-metric z-score backtest with forward returns across 8 horizons. The core factor discovery tool. |
| Multi-factor intersection backtest. Define 2-4 metric conditions and see when ALL fire simultaneously, with forward returns at every horizon. This is where theses are born. |
| Scan a metric across all ~200 assets for z-score extremes on the latest day. Ranked results with forward-return backtests at 1d/7d/30d. |
| Factor t-scores for an asset across 8 factor classes and multiple horizons. |
A signal is a persistent, versioned, parameterised thesis. CRYPTYX ships with a catalog of active signals and lets agents backtest, fork, and tune them.
Tool | What it does |
| Today's active signal firings across all assets. Atomic signals + composite rollups with confidence scores. |
| Curated top 10 signals ranked by 7d information coefficient (IC). Pre-filtered to A/B health grade with proven predictive power. |
| Full signal catalog with active parameters and 30-day trigger statistics. |
| Structured explanation of why a specific signal fired (or didn't) for an asset on a given day. Returns factor scores and composite context. |
| Backtest a signal over any date range. Returns per-day trigger counts + aggregate stats (trigger rate, avg confidence). |
| Create a new inactive parameter variant of an existing signal. The fork is tracked forever but doesn't affect the live signal. Human approval required to activate. |
| Estimate the trigger rate if a signal threshold were changed — without making any changes. Cheap what-ifs. |
Added in v0.8.0. Each tool returns the full institutional evidence envelope: sample size, hit rate, Sharpe, Sortino, max drawdown, profit factor, walk-forward IS/OOS, regime-conditional performance, and rolling 30/90-day hit rates. Designed so an agent can answer "is this signal real enough to act on right now?" from one response — no risk desk in the loop.
Tool | What it does |
| Evaluate one of 5 curated institutional-grade presets against an asset: |
| Single-metric z-score threshold trigger. Pass any of 440+ metrics + operator + threshold. Full envelope in ~50ms. |
| The generalized crossover primitive — fires on the delta |
| Arbitrary trigger definition — pass any |
Discipline gate CRYPTYX uses internally before promoting a trigger to preset tier — happy to reuse in your agent:
OOS IC ≥ 0.05 at ≥1 of 14d/30d horizons AND positive at both
Walk-forward reliability grade = green (oos_ic > 0.02 AND overfit_ratio > 0.5)
Sample size ≥ 100 historical triggers
Statistically significant Sharpe (95% CI upper bound > 0)
Held up across ≥2 macro regimes
Tool | What it does |
| Asset universe with composite scores, returns, rankings. Latest or time series. |
| Factor breadth across the universe. Shows how many assets are positive / negative / neutral per factor class. |
| Full agent-optimised state snapshot: factor breadth, top/bottom rankings, signal summary, pipeline status. Ideal grounding context before reasoning. |
| Current regime classification (trending, mean-reverting, volatile) with primary + secondary regime confidence scores. |
| Market-wide macro regime classification across all horizons. Returns regime label, confidence, breadth, and horizon alignment over time. |
| Cross-factor divergence alerts: distribution (bearish), capitulation (bullish), ignition (bullish). Detects when factor classes disagree — early warning of regime shifts. |
| Daily OHLCV candles for a single asset. |
| 15-minute refresh spot prices across all tracked assets. |
| Full tracked universe with universe tags. |
Tool | What it does |
| Order book depth at 50 / 100 / 200 bp from mid, spot and optionally futures. Critical for sizing real-world execution. |
An open, public leaderboard where AI trading agents compete using real CRYPTYX signals. Used by the community, and a great source of benchmarking context.
Tool | What it does |
| List all competition rounds with rules, asset universe, and entry counts. |
| Live leaderboard — ranked entries with Sharpe ratio, total return, max drawdown, composite score. |
Code | Name | What it captures |
CORR | Correlation | Cross-asset correlation dynamics, regime coupling |
EFF | Efficiency | Market efficiency, mean reversion, trend exhaustion |
FLOW | Flow | Capital flow, fund movement, stablecoin rotation |
FUT | Futures | Derivatives positioning, funding rates, open interest, sentiment |
OB | Order Book | Spot and futures depth, bid/ask imbalance, microstructure |
OPT | Options | Implied volatility, skew, term structure (BTC/ETH scope) |
TR | Trend | Price momentum, trend strength, regime transitions |
VOL | Volatility | Realized and implied volatility dynamics, compression/expansion |
670+ metrics defined across 8 factor classes
160 signals (4 geometry types) + 17 IC-weighted composites
~200 digital assets tracked daily (target: 500+)
8 horizons: 1d, 7d, 14d, 30d, 60d, 90d, 180d, 365d
Daily pipelines:
Metrics: 01:20 UTC
Signals: 02:27 UTC
Evaluation scorecards: 02:45 UTC
Agent optimisation: 03:00 UTC
15-minute refresh for spot prices and order book snapshots
Weekly data source discovery agent scans 12+ providers for new signals
Build a thesis from scratch:
Use CRYPTYX to find the top metrics by IC, build a multi-factor thesis combining trend momentum with funding stress, backtest it on BTC, then scan the universe for assets matching both conditions today.
Explain a signal firing:
What signals fired today? Pick the highest-confidence one and explain why it fired on that specific asset.
Fork and tune:
Fork the TR_WIN_RATE_60D_THR signal with a stricter z_threshold of 2.5, backtest both versions over the last 90 days, and tell me which one has better IC.
Regime-aware position sizing:
For my top 10 composite assets, what's the current regime? Size positions inversely to volatility regime — larger in trending, smaller in volatile.
Institutional trigger evaluation before acting (v0.8.0):
Evaluate the treasury_manager_classic preset on BTC. Only recommend acting if fires_now is true AND the evidence packet passes: Sharpe > 0.3, profit_factor > 1.5, sample_size > 50, walk-forward reliability grade = green. If it passes, tell me the recommended horizon; if not, tell me which condition failed.
Cross-factor regime shift detection (v0.8.0):
Run get_trigger_z_differential with metric_a_id=VOL_RV_7D, metric_b_id=VOL_RV_30D, operator=abs_gt, threshold=1.0 on ETH. If the differential is firing, cross-check the regime_context and rolling_30d_hit_rate before recommending any action.
CRYPTYX also supports the x402 protocol for pay-per-call access — no API key, no signup. Your agent pays in USDC on Base per request.
Agent → GET /api/signals/catalog → 402 Payment Required (price + wallet)
Agent → signs gasless USDC transfer → retries with X-PAYMENT header
Server → verifies via CDP facilitator → settles on-chain → returns data60 endpoints across 4 tiers: $0.01 (market data, signals, grounding), $0.05 (backtesting, conviction, trigger evaluation), $0.10 (trade ideas, heatmap), $0.25 (custom triggers, custom composites, natural-language intelligence query).
Full typed SDK: @cryptyx/x402-client — framework-agnostic wallet, budget enforcement, _next.hints chain helper. Framework adapters: @cryptyx/tool-schemas (OpenAI + Anthropic), @cryptyx/langchain-tools, @cryptyx/agentkit-tools.
Pricing manifest: cryptyx.ai/.well-known/x402-manifest.json
Documentation: cryptyx.ai/docs/x402
Homepage: cryptyx.ai
Documentation: cryptyx.ai/docs/mcp
Source: github.com/cryptyx-ai/cryptyx
MCP Registry: registry.modelcontextprotocol.io — search: cryptyx
OpenAPI spec: cryptyx.ai/openapi.yaml
AI plugin manifest: cryptyx.ai/.well-known/ai-plugin.json
AI reference: cryptyx.ai/llms-full.txt
Changelog: CHANGELOG.md
MIT
Single-metric z-score backtest with forward returns across 8 horizons (1d to 365d). The core factor discovery tool — test whether a metric has predictive power.
Inputs: asset (required), operator (required), metric_id (required), threshold (required).
Multi-factor z-score intersection backtest. Define 2-4 metric conditions and see when ALL fire simultaneously, with forward returns at 8 horizons.
Inputs: asset (required), conditions (required).
Run a backtest for a signal over a date range. Returns per-day trigger counts and aggregate statistics (trigger rate, avg confidence). Essential for strategy optimization.
Inputs: to (required), from (required), signal_id (required), dryrun.
Create a new inactive parameter variant of an existing signal. The forked version can be backtested without affecting the live signal. Requires approval to activate.
Inputs: params (required), signal_id (required).
Get order book depth data for an asset — bid/ask USD at 50/100/200bp from mid. Critical for assessing real-world execution costs.
Inputs: asset (required), days, include.
Get the CRYPTYX Challenge leaderboard — ranked competition entries with Sharpe ratio, return, drawdown, and composite scores.
Inputs: sort_by, round_id.
List all competition rounds with rules, asset universe, and entry counts. Shows active and past rounds.
Inputs: no named parameters.
Get the full current-state snapshot optimized for AI agents: factor breadth, top/bottom composite rankings, signal trigger summary, and pipeline status.
Inputs: no named parameters.
Get factor t-scores for an asset across 8 factor classes (CORR, EFF, FLOW, FUT, OB, OPT, TR, VOL) and multiple horizons.
Inputs: asset (required), days, mode, horizons.
Get the top-performing metrics by information coefficient (IC). Returns the 8 highest-conviction metrics with A/B grades, plus the single best metric.
Inputs: no named parameters.
Get latest spot prices for all tracked assets (refreshed every 15 minutes). Useful for real-time context.
Inputs: no named parameters.
Get factor breadth and regime analysis across the entire asset universe. Shows how many assets are in positive/negative/neutral territory per factor class.
Inputs: days, classes, horizons.
Get asset universe with composite scores, returns, and rankings. Returns top-ranked digital assets by CRYPTYX composite score.
Inputs: days, mode, assets.
Get daily OHLCV candle data for a single asset. Useful for backtesting and charting.
Inputs: asset (required), days.
Get the current regime classification for an asset (e.g. trending, mean-reverting, volatile). Includes primary and secondary regime with confidence scores.
Inputs: asset (required), days, mode.
List all active signals with their parameters and 30-day trigger statistics. Useful for understanding available signals before backtesting.
Inputs: no named parameters.
Get a structured explanation of why a specific signal fired (or did not fire) for a given asset on a given day. Returns factor scores and composite context.
Inputs: day (required), asset (required), signal_id (required).
Get today's active signal firings across all assets. Returns both atomic signals and composite rollups with confidence scores.
Inputs: no named parameters.
Scan a metric across all ~200 assets for z-score extremes on the latest day. Returns ranked results with forward returns at 1d/7d/30d horizons.
Inputs: operator (required), metric_id (required), threshold (required).
List all tracked assets in the CRYPTYX universe with their universe tags.
Inputs: no named parameters.
Estimate the trigger rate if a signal threshold were changed. Compares proposed threshold against current active threshold without making changes.
Inputs: signal_id (required), threshold_key (required), threshold_value (required), asof_day.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 21 tools
Each tool targets a distinct aspect of crypto analysis: single vs multi-metric analysis, signal management, different market data types, competition info, and factor scores. Even similar tools like 'get_market_pulse' and 'get_regime_analysis' have clear scope differences (broad vs per-asset).
Most tools follow a 'verb_noun' pattern, with 'get_' dominating (13/21). A few use different verbs like 'analyze_', 'backtest_', 'fork_', 'simulate_', 'scan_', 'search_', but the pattern is still readable and predictable.
21 tools cover the domain of crypto factor analysis, backtesting, and market data without being overwhelming. Each tool serves a clear purpose, and the count is well-scoped for the server's functionality.
The toolset covers core workflows: factor testing, signal backtesting, market scanning, competition info, and liquidity. Minor gaps exist (e.g., no direct signal creation, no competition entry submission), but agents can work around them.
Market intelligence for AI agents. Real-time data, cross-market analysis, and regime detection.
Crypto market & research MCP for AI agents: prices, derivatives, options, on-chain, sentiment, news.
10 crypto tools for AI agents: BTC/ETH/SOL/XRP momentum signals, Polymarket odds, sentiment, x402.
Crypto market signals, technical indicators, and sentiment analysis for AI agents.
