Shannon Entropy Calculator
This tool computes Shannon entropy right in your browser — the text never leaves the page. Three alphabets are available: Unicode characters, UTF-8 bytes and words. Besides the whole-text summary there is a per-line mode: lines with unusually high entropy get highlighted — a handy way to spot tokens, keys and passwords that leaked into logs and config files.
What Shannon entropy is. Entropy H = −Σ pᵢ·log₂pᵢ is the average amount of information (in bits) per symbol of a message. The more diverse and uniform the symbol distribution, the higher the entropy: a string of identical letters scores zero, a uniform alphabet of N symbols scores log₂N. Meaningful text is redundant and yields 4–4.5 bits per character, while random data (keys, tokens, compressed files) approaches the maximum — which is exactly why entropy works for secret scanning.
Text summary
| Entropy H: | — |
| Alphabet size: | — |
| Normalized entropy H/H₍max₎: | — |
| Redundancy: | — |
| Total units: | — |
| Information H×N: | — |
| Theoretical compression limit: | — |
Reference scale (bits/unit)
Frequency table
| Unit | Count | Probability | Contribution, bits |
|---|
Per-line analysis
| # | Entropy | Length | Content |
|---|
Free online Shannon entropy calculator. Computes text entropy over Unicode characters, UTF-8 bytes or words, and reports alphabet size, normalized entropy, redundancy and the theoretical compression limit. The frequency table breaks down each symbol's contribution to the total.
The per-line mode is built for secret scanning: lines of a log or config file whose entropy exceeds an adjustable threshold get highlighted as candidates — that is how accidentally committed API keys, tokens and passwords are found. The same technique powers secret scanners like truffleHog and gitleaks.
Everything runs entirely in your browser: the text is never sent to a server, so the tool is safe to use with private data. A reference scale of typical values is provided for comparison — from meaningful text (~4 bits per character) to random bytes (8 bits).