Whitepaper

Version 1.0 — Published 2026-05-10

Abstract

Razalith is a continuous cryptocurrency anomaly detection platform that assigns a 0–100 anomaly score to each of 250+ digital assets. Every score measures how unusual an asset’s recent behavior is compared exclusively to its own historical patterns — never to the broader market, other coins, or any external benchmark.

The platform delivers email alerts, Telegram alerts for Pro users, AI-powered per-asset market analysis, and a fully transparent alert track record where every result is public. Razalithis an informational analytics tool — it does not provide financial, investment, or trading advice.

Razalith has never issued any cryptocurrency token, coin, or digital asset and has no plans to do so. Any token using the Razalith name is fraudulent.

1. Introduction

1.1 The Problem

Cryptocurrency markets run 24/7 across hundreds of assets with extreme volatility. Traders, investors, and analysts face the same challenge: spotting genuinely unusual activity early while filtering out normal fluctuations and market-wide noise.

Most existing tools compare assets against each other or against aggregate benchmarks. This conflates asset-specific anomalies with broad market movements. A coin that drops 5% during a market-wide correction is behaving normally; the same 5% drop during a calm market may be highly significant.

1.2 The Razalith Approach

Razalith takes a fundamentally different path: self-referential scoring. Each asset is compared only to its own historical behavior. This isolates genuinely unusual activity specific to each asset, regardless of what the broader market is doing.

2. Scoring Methodology

2.1 Self-Referential Scoring

The core principle behind Razalithis self-referential analysis. Every anomaly score is computed by comparing an asset’s current state against that same asset’s own recent history:

• Bitcoin is only compared to Bitcoin’s past behavior
• Ethereum is only compared to Ethereum’s past behavior
• No asset is ever benchmarked against any other asset or market index

2.2 Proprietary Multi-Factor Model

Each anomaly score (0–100) is generated by Razalith’s proprietary multi-factor model. The model evaluates multiple independent dimensions of market behavior, all measured against that specific asset’s own baseline. The exact dimensions evaluated, their composition, weights, and calibration are proprietary.

The model is designed so that multiple independent signals must converge before a high score is produced. A single unusual reading in one dimension alone is not enough to generate an extreme score — it takes agreement across the model’s factors.

2.3 Score Interpretation

A high anomaly score is non-directional: it flags that something unusual is happening but does not predict whether the price will move up or down. The score measures unusualness, not direction.

2.4 Market Bias Indicator

Independently of the anomaly score, Razalith computes a directional market bias for each asset: Bullish, Bearish, or Neutral. The bias is derived from the convergence of multiple directional signals and is displayed alongside the anomaly score. It has no influence on the anomaly score itself.

2.5 Data Quality Safeguards

Razalith enforces strict data quality controls to prevent false or misleading signals:

• Freshness enforcement— Scores are automatically capped when the underlying price data is not recent enough, preventing stale signals from reaching users.
• Low-activity filtering— Assets with abnormally low trading activity relative to their own history are scored more conservatively to prevent noise from thin markets.
• Incomplete data handling— Assets with insufficient historical data for a full multi-factor evaluation are constrained to lower score ranges until enough context is available.
• Continuous calibration— Internal monitoring ensures that score distributions across the asset universe remain within expected statistical bounds. Deviations trigger system-level reviews.

3. Continuous Scanning

3.1 Data Pipeline

Razalith continuously monitors the top 250+ cryptocurrencies by market capitalization. Price and volume data is sourced from established third-party market data aggregators that compile data from major exchanges worldwide.

3.2 Scan Frequency

• Pro users — Full anomaly scans every 15 minutes for rapid detection.
• Free users — Anomaly scans every 60 minutes on the top 25 assets.

3.3 Historical Context

Each scoring cycle draws on an extended window of historical price data to establish each asset’s behavioral baseline. This rolling window ensures the model adapts to evolving market conditions while retaining enough context to identify genuinely unusual events.

4. Alert System

4.1 Alert Triggers

When an asset’s anomaly score reaches 75 or higher (Extreme level), Razalith automatically fires an alert. Users can customize their alert threshold between 50 and 100 to match their desired sensitivity.

4.2 Delivery Channels

• Telegram (Pro only) — Instant push notifications via the Razalith Telegram bot with one-time account linking.
• Email— Detailed alert emails with the anomaly score, market bias, price context, and a direct link to the asset on the dashboard.

4.3 Cooldown

To prevent notification fatigue, Razalith enforces a cooldown period per asset after each alert. Subsequent extreme readings during the cooldown are recorded internally but do not trigger duplicate notifications.

5. AI-Powered Market Analysis

Razalith provides on-demand AI-generated analysis for each monitored asset. When a user requests an analysis, the system assembles the asset’s current market context and submits it to an advanced language model for synthesis into a plain-language narrative.

Every AI analysis follows strict guardrails:

• Uses only factual data from Razalith’s own scoring system
• Never provides price predictions or investment recommendations
• Explains what is happening and why it is notable, in clear language
• Identifies whether signals agree or conflict with each other
• Highlights the single most important factor to watch

Free users receive 5 analyses per day. Pro users receive 250 per day.

6. Track Record & Transparency

6.1 Why a Public Track Record Matters

Most analytics platforms make claims about their accuracy without any verifiable evidence. Razalith takes the opposite approach: every single alert is publicly tracked and measured against real market outcomes.

6.2 How It Works

Every extreme alert (score ≥75) is automatically added to Razalith’s public Track Record. For each alert, Razalith records:

• The asset name and symbol
• The exact date and time of the alert
• The anomaly score at alert time
• The maximum price move (up or down) within 7 days of the alert

6.3 Success Measurement

An alert is classified as a “success” if the asset moved at least ±5% from the alert price within the 7-day evaluation window. This metric is non-directional — it measures whether the anomaly correctly identified a significant move, regardless of whether that move was up or down.

6.4 No Cherry-Picking

All alerts are tracked automatically. No results are hidden, filtered, or selectively displayed. The track record includes both successful and unsuccessful alerts. Pending alerts still within the 7-day window are clearly marked as such. Anyone can verify the results at any time.

7. Regime Detection

Beyond the composite anomaly score, Razalithidentifies notable live market conditions for each asset. These “regime signals” flag distinct structural situations — moments when the character of the market has shifted in a way that warrants attention, independent of the anomaly score itself.

The specific conditions detected and the methods used to identify them are proprietary. Regime signals are surfaced in the AI analysis and alert context to give users richer situational awareness without requiring manual chart interpretation.

8. Plans & Pricing

Full pricing details are available on the Pricing page. Cancel anytime. No hidden fees.

9. Infrastructure & Security

Razalith runs on modern cloud infrastructure with globally distributed edge computing for low latency. All data in transit is encrypted via TLS. User authentication and session management are handled by enterprise-grade identity providers. Market data is sourced from established aggregators that compile data from major global exchanges.

Razalith does not operate an exchange, hold funds, or provide custodial services of any kind.

10. What Razalith Is Not

• Not a trading platform — You cannot buy, sell, or hold any asset on Razalith.
• Not financial advice— Anomaly scores, AI analyses, and all content are purely informational.
• Not a price predictor— High scores indicate unusualness, not price direction.
• Not a custodial service — Razalith does not hold, transfer, or manage any funds or digital assets.
• Not a token issuer — Razalith has never issued any cryptocurrency token, coin, NFT, or digital asset. Any token using the Razalith name is fraudulent.

11. Further Reading

For accessible introductions to the concepts behind Razalith, visit the Learn section:

• What Is Crypto Anomaly Detection?
• Self-Referential Scoring: Why It Matters
• Track Record Methodology
• How Crypto Anomaly Alerts Work
• Anomaly Detection vs. Technical Analysis

12. Roadmap

Razalith is under active development. Planned improvements include expanded asset coverage, next-generation scoring models, enhanced regime detection, portfolio-level anomaly monitoring, API access for developers, and deeper historical analytics. Updates ship continuously.

13. Legal & Disclaimers

Razalith provides market data and anomaly scores for informational purposes only. Nothing on this platform constitutes financial, investment, trading, tax, or any other form of professional advice. Digital asset markets are extremely volatile and carry a high risk of capital loss. All data is provided “as is” without warranty. AI-generated analyses may contain errors or inaccuracies.

For full legal terms, see the Terms of Service, Privacy Policy, and Disclaimer.

14. Contact

General inquiries: hello@razalith.com

Follow Razalith: X (Twitter) · Telegram