Skip to content

Text Entropy Calculator

Calculate empirical Shannon entropy from scalar, grapheme or UTF-8 byte frequencies. Inspect probabilities and export a report without security claims.

Choose your path

Built around the job you need to finish

Measure empirical scalar, grapheme or UTF-8 byte distributions, expose probabilities and Shannon contributions, and separate evenness from unpredictability.

Information-theory learner

Reproduce entropy from individual probabilities.

Compare AAAA, ABAB, AABB and AAAB and add the displayed contributions.

Understands zero entropy for one symbol and equal entropy for reordered distributions.

Encoding analyst

Distinguish text-symbol entropy from encoded-byte entropy.

Analyze one emoji as scalars then UTF-8 bytes; inspect byte identities and normalization choices.

Knows the selected observation unit and does not call it password strength.

Mobile dataset reviewer

Keep a complete, inspectable analysis record.

Import a fixture, change policies, review long distributions and export the complete report.

Old results are invalidated and model assumptions travel with the report.

Was this tool helpful?

Reference & details

How it works

Observed distribution

For each selected symbol, p=count/N and H=−Σp·log2(p). Base 2 gives bits per symbol. Empty input returns zero by convention.

Observed alphabet

The maximum log2(k) uses observed distinct symbols only. Evenness H/log2(k) is undefined for fewer than two symbol types.

Interpretation limits

Symbol frequencies ignore order and dependence. Evenness and N×H do not estimate passwords, cryptographic randomness, prediction difficulty or actual compression size.

Updated: August 2026

Example Scenarios

Reproduce entropy from individual probabilities. Compare AAAA, ABAB, AABB and AAAB and add the displayed contributions.

Distinguish text-symbol entropy from encoded-byte entropy. Analyze one emoji as scalars then UTF-8 bytes; inspect byte identities and normalization choices.

Keep a complete, inspectable analysis record. Import a fixture, change policies, review long distributions and export the complete report.

Common Mistakes to Avoid

Calling uniform frequency random

A predictable repeating sequence can have uniform symbol frequency.

Reading N×H as actual compressed size

It is an independent-symbol model quantity, not a compressor or file-format measurement.

FAQ

Both have the same symbol counts. This empirical distribution measure ignores sequence order.

In scalar mode, a sequence containing only one emoji scalar has zero entropy. UTF-8 byte mode is a different distribution.

No. The generator, dependencies, attacker knowledge and guessing model are not measured here.

It is log2(k) for the alphabet observed in this sample, not the maximum of an external alphabet or a generating process.

All frequency rows, probabilities, contributions, settings and limits are included in a JSON report. Preview paging does not truncate it.

About Text Entropy Calculator

Measure empirical scalar, grapheme or UTF-8 byte distributions, expose probabilities and Shannon contributions, and separate evenness from unpredictability.