Is your crypto safe?

Trust ratings for 16,300 tokens. Crash predictions. DeFi yield safety.

Safest crypto · Crash watch · Yield safety · Contagion map

We predicted 144 crypto crashes with 98% precision · 22-month avg lead time

Crypto Intelligence Platform

Independent ratings.
Distance-to-Default.
Before the crash.

Institutional-grade crypto intelligence. We rate 15,000+ tokens, 1,000+ exchanges, and 7,000+ DeFi protocols. Trust ratings on a Aaa–D scale, Distance-to-Default scoring, structural collapse detection, and conviction-ranked trading signals. API-first. Machine-readable. Free during beta.

100%
Defaults Detected
98%
Precision

Out-of-sample validation (Jan 2024 – Feb 2026): 113 of 113 token collapses detected. 98% of flagged tokens lost >50% of their value. Methodology →

Live Structural Alerts
View all 47 alerts →
Out-of-sample: 176 collapse signals issued, 98% lost >50%. Only 1 false positive (<30% decline).
Live Signal Accuracy 47 active alerts
91%
1st Target Hit
41 of 45 mature alerts declined ≥30% from alert price
Latest: XPL crossed -30% on 2026-02-05
84%
2nd Target Hit
38 of 45 mature alerts declined ≥50% from alert price
Latest: B crossed -50% on 2026-04-16
47
Active Alerts
7 Collapse · 40 Stress
2
Monitoring
Recent alerts, tracking
LION
Latest Alert Issued
Stress · 2026-02-09
Latest to hit target: B reached -50% on 2026-04-16
Portable Alpha — Backtest Track Record (Apr 2021 – Dec 2025)
Conservative
+314.7%
vs BTC +100.8% · Alpha +213.9%
2.02
Sharpe
-11.0%
Max Drawdown
75.0%
Win Rate
24
Months OOS
Growth
+842.2%
vs BTC +100.8% · Alpha +741.4%
3.86
Sharpe
-12.1%
Max Drawdown
75.0%
Win Rate
24
Months OOS
Aggressive
+1,872.9%
vs BTC +100.8% · Alpha +1,772.1%
5.56
Sharpe
-18.8%
Max Drawdown
75.0%
Win Rate
24
Months OOS
Full track record →
Agent Intelligence — Live
47,119 AI agents tracked across 9 chains
LIVE
Capital at Crash Risk
$517M
crash probability >50%
Structural Collapse
12 agents
$344M exposed · 90% crash prob
Base Chain Risk
19,442 agents
$1.92B · concentration risk 10/10
Latest finding: Agent pippin (Virtuals) holds $344.6M in PIPPIN token, currently rated CRITICAL with 90% crash probability and structural_weakness=3 — the highest possible collapse signal.
Full Agent Intelligence report →
Yield Risk Intelligence — Live
59 active divergence alerts across 6,539 yield pools
LIVE
Active Alerts
59
TVL exodus · APY spikes · cliffs
Yield Traps
2,553
27 EXTREME risk
Pools Monitored
6,539
18,784 indexed · 119 chains
TVL Covered
$140B
across 513 protocols
View all yield risk alerts →
What We Do
1
Crash Prediction
Probabilistic crash forecasting combining rating trajectories, DtD trends, TVL divergence, and DeFi protocol exposure. Our model detects “looks healthy but dying” scenarios that surface-level metrics miss.
98% precision
2
Distance-to-Default
Our DtD model measures how far each token is from critical failure — using on-chain metrics, liquidity signals, and historical default patterns. Tokens with DtD below 2.0 are in the danger zone. Based on the same framework credit agencies use for bonds, adapted for digital assets. Average detection lead time: 22 months before default.
100% recall on defaults
3
Portfolio Stress Testing
Stress-test any portfolio against 5 historical crises: FTX collapse, LUNA death spiral, 3AC lending contagion, Oct 2025 flash crash, and BTC bear market. Calibrated against actual price data — 82% of predictions within 10 percentage points of the real crash. SOL predicted -67.6% vs actual -67.9% during FTX. Median error: 3.4pp across 67 token-scenario pairs.
3.4pp median error →
4
Contagion Map
Interactive network graph of 198 tokens showing how ecosystems connect and where risk concentrates. Run crisis simulations to see how a failure in one token cascades through the market. Uses the same calibrated model as the stress tester — validated against FTX, LUNA, 3AC, and Oct 2025 flash crash with 96% of predictions within 15pp.
198 tokens mapped →
5
Portable Alpha Strategy
Conviction-ranked long/short pairs generated from rating spreads and DtD differentials within investment-grade tokens. Backtested April 2021 — December 2025: Sharpe ratio 2.82. Paper trading live since March 2026, SHA-256 audited.
Paper trading live
6
Trust Ratings
Every token scored 0–100 across five pillars: Ecosystem Health, Contagion Risk, Historical Resilience, Fundamental Quality, and Rug Pull Resistance. Mapped to a Moody’s-style letter scale from Aaa to D. Updated monthly.
210 tokens rated
7
Cascade Risk Engine
Simulate how a token collapse propagates through the crypto ecosystem. ZARQ’s in-memory graph maps 210+ tokens across dependency chains — ecosystem, correlation, and bridge links. Run any scenario in under 100ms: structural collapse, BTC crash, stablecoin depeg, or exchange failure. Identify systemic hotspots and stress-test your portfolio against cascade contagion.
210 tokens mapped →
8
Machine-First API
Every data point accessible via REST API. Structured for LLM consumption, AI agent integration, and automated decision pipelines. Rate-limited to 1,000 calls/day during beta. No API key required.
Free during beta
9
Agent Intelligence
The only platform that crosses 47,119 AI agents with institutional crypto risk data. We track which agents hold which tokens, flag exposure to Structural Collapse and CRITICAL-rated assets, and rank chains by AI-agent concentration risk. $517M in agent capital currently exposed to tokens with >50% crash probability. Agent Exodus Index — detecting when agents leave before TVL collapses — launching Q3 2026.
47,119 agents tracked →
API

Built for machines.
Used by humans.

Every rating, signal, and prediction available via REST. Structured JSON responses designed for LLM context windows, trading bots, and risk dashboards. No authentication required during beta.

GET /v1/crypto/rating/bitcoin { "token": "bitcoin", "rating": "A2", "score": 74.52, "pillars": { "ecosystem": 99.20, "contagion": 32.84, "resilience": 71.59, "fundamental": 86.50, "rug_pull": 94.75 }, "dtd": 3.03, "risk_level": "WARNING" }
Product Suite
ZARQ Signal
Live risk feed for 205 tokens. Filterable by severity. Auto-refreshes every 60s.
/signal →
Know Your Agent
Due diligence on 204K AI agents & tools. Trust score, risk level, compliance status.
/kya →
Save Simulator
See the crashes ZARQ predicted. Real out-of-sample saves an agent would have made.
/demo/save-simulator →
API Documentation
Full endpoint reference. Python, JS, and cURL examples. Free during beta.
/docs →
For Developers

Add crypto risk checks to your pipeline. Free, no auth required.

curl https://zarq.ai/v1/check?token=bitcoin

# {"rating": "A2", "trust_score": 85, "crash_probability": 12.3, ...}
API docs try check API try preflight

Free during founding period.

All endpoints open. 1,000 calls/day. No API key required.

Explore the API

Disclaimer — ZARQ provides data-driven crypto intelligence for informational and educational purposes only. Nothing on this website constitutes financial advice, investment advice, trading advice, or any other form of professional advice. ZARQ does not recommend buying, selling, or holding any cryptocurrency or financial instrument.

Trust ratings, Distance-to-Default (DtD) scores, crash probabilities, trading signals, and all other data are derived from quantitative models and may contain errors, lag behind real-time conditions, or fail to predict future outcomes. Past performance, whether backtested or live, is not indicative of future results. All backtested results are hypothetical and do not represent actual trading.

Crypto assets are highly volatile, speculative, and may result in total loss of invested capital. You should conduct your own research (DYOR) and consult a qualified, licensed financial advisor before making any investment decisions. ZARQ is not registered as a broker-dealer, investment advisor, or financial institution in any jurisdiction.

By using this website, you acknowledge that you bear sole responsibility for your own investment decisions and that ZARQ, its founders, contributors, and affiliates accept no liability for any losses, damages, or consequences arising from the use of information provided herein.