entviz¶

entviz turns a high-entropy value — a cryptographic key or signature, a UUID, a blockchain address, a post-quantum key, a genome — into a compact SVG diagram that an ordinary person with reasonably good vision can compare at a glance. The goal is simple: let a human decide "are these two blobs of entropy the same or different?" without reading 88 characters of base64 one symbol at a time.

Every value is rendered across several redundant visual channels — text cells, a per-cell surround pattern, nucleus colors, a fingerprint-derived color bar, blank-cell markers, and an ellipse overlay — so that even a single-bit difference in the input is obvious, and so that color-blind viewers and monochrome displays still get a reliable signal.

Explore¶

• Specification — the full algorithm (current: v15): normalization, tokenization, the fingerprint, geometry, and every visual channel.
• Gallery — entvizes across real input types (UUIDs, hex, blockchain addresses, SSH keys, ULIDs, LEIs, snowflakes) and avalanche pairs.
• Paper — the longer-form analysis and design rationale.
• Developers (README on GitHub) — install with uv, run the CLI, run the tests, and cut releases.

A threat model covering the comparison guarantees is also available.

Implementations¶

entviz is a language-independent specification; every implementation below passes the same shared conformance corpus. ▶ Try it live in the browser — the JavaScript/React playground lets you render a value and compare two.

At a glance¶

• Losslessly represents up to 512 bits in the text channel; larger inputs show head + a domain-separated fingerprint readout in the middle + tail, and bind the whole input through the fingerprint.
• Amplifies single-bit differences via a SHA-512 fingerprint, even when the input itself has no avalanche effect.
• Usable under red-green, blue-yellow, and complete color blindness.
• Trivial to implement, with no significant dependencies.