엔트로피 시각화 도구

(문서는 영어)

What it does

The

EntropyEntropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

Visualizer measures how random your data really is, byte by byte. Paste text or drop any file and it computes Shannon

entropyentropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

(bits per byte, 0-8), a byte frequency histogram (256 bins), a chi-squared uniformity test with p-value, the serial correlation between consecutive bytes, and a Monte Carlo π estimate — the five classic randomness checks from the ent test suite. A block

entropyentropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

heatmap shows where in the data the predictable stretches hide.

This matters wherever randomness is a security property. Encryption keys, API tokens, session IDs, and password salts are only as strong as the

entropyentropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

behind them: a “random-looking” key generated by a weak source can be brute-forced orders of magnitude faster than its length suggests. Feed the tool your generator’s output and the verdict card tells you immediately whether it is excellent (indistinguishable from noise) or leaking structure. Everything runs 100% client-side — your data never leaves the browser.

How to use it

1. Paste text into Data to analyze, or drop any file onto the box (or click Open file). Analysis runs automatically, ~250 ms after you stop typing.
2. Read the verdict badge at the top: excellent, good, suspicious, or low.
3. Check the four metric cards: the Shannon

   entropyentropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

   gauge (0-8 bits/byte), chi-squared with pass/fail, Monte Carlo π vs the real π, and serial correlation (0 = independent bytes).
4. Scan the byte frequency histogram: green bars hug the dashed expected-count line; red bars are over- or under-represented bytes.
5. Read the

   entropyentropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

   heatmap: one cell per 16-byte block, red = predictable, blue = high

   entropyentropyThe unpredictability of a secret, measured in bits. Each added bit doubles the guesses an attacker needs; strength comes from length and randomness, not from obscurity.

   — spots of red inside blue data are embedded patterns.
6. Copy report for a plain-text summary, or Copy share link to reproduce the exact view (text inputs are encoded in the URL).

Examples

Repeating text is not random

Input: abcabcabcabc… (900 bytes)

Shannon entropy: 1.585 bits/byte (low)   — only 3 distinct bytes
Chi-squared: 75900.0 (p=0.0000) — skewed — fail
Serial correlation: -0.5000

Good generator output looks like noise

Input: 65,536 bytes from a seeded PRNG (fixed seed 42, mulberry32)

Shannon entropy: 7.9972 bits/byte (excellent)
Chi-squared: 252.1 (df 255, p=0.5389) — uniform — pass
Serial correlation: -0.0083
Monte Carlo π: 3.1272 (actual 3.14159)

English text sits in between

Input: The quick brown fox jumps over the lazy dog. 0123456789 abcdef ABCDEF

Shannon entropy: ≈ 5.1 bits/byte (suspicious)
Histogram: ASCII letters spike green-red; bytes 128-255 nearly absent

Good to know

• Verdict scale: > 7.5 bits/byte = excellent, > 6 = good, > 4 = suspicious, ≤ 4 = low. Compressed and encrypted data should always land near 8.
• Small samples lie: chi-squared and Monte Carlo π need thousands of bytes to be meaningful. Short inputs show noisy numbers — that is expected, not a failure.
• Files over 512 KB are analyzed up to that limit (the header line shows when truncation happened); the heatmap shows at most 4,096 blocks.
• Privacy: all analysis is local — nothing is uploaded, no cookies, no accounts. Safe for keys and secrets.
• Related tools: Password Strength Analyser, Secure Token Generator,

  HashHashA one-way function that maps data of any size to a fixed-length digest. Same input, same digest; any change, different digest; not reversible.

  .