Drop a .solus file. The full audit — cryptographic provenance and the writer's typing pattern — runs entirely in your browser. Nothing is uploaded.
or click to browse your computer
Works with all Solus documents. No account required.
Solus doesn't hand out a single "authenticity score." A .solus file's cryptographic provenance and its writer's typing pattern are graded independently, then combined into a verdict. Both run locally — your file never leaves this page.
Purely cryptographic. The signature has to verify against the certificate body, the certificate body's hash has to match the commitment, the ledger's Merkle root has to reproduce, the paste-policy log has to match the certificate's claim.
20 checks · fail any one → verdict is tampered
Behavioral. Does the writer have an established typing baseline, and does this submission's rhythm match it? Composition-vs-transcription signal, pause-location distribution, keystroke sufficiency, anti-transcription composite — all measured against the writer's own history.
10 checks · fail any hard one → unsupported; partial → limited
Axis A is available immediately — all 20 cryptographic checks run on the very first document anyone writes, with no history and no account. Axis B is different: it compares a submission against that writer's own baseline, and a baseline has to be built out of their real writing first. Until it exists, the Axis B checks report amber. That amber describes missing evidence, not suspicious evidence.
For most writers that's the first two or three assignments. Verify any file from an enrolling writer and the result shows live meters for exactly how far along they are and what's still outstanding — so "not yet" always comes with a finish line. Once the baseline completes, Axis B runs on every subsequent submission and Verified Human™ becomes available.
Five possible outcomes. The badge only appears for the first one.
Provenance intact and the writer's typing pattern matches their enrolled baseline. The strongest claim Solus can make.
Cryptographic checks pass, but one or more behavioral signals came back mixed. Not a tamper; teacher reviews in context.
File is cryptographically clean, but this writer has no typing baseline to compare against, so the behavioral axis couldn't run. Normal for a new writer's early work and for legacy files — an absence of evidence, not evidence against. Machine token: unsupported
Provenance intact, and the writer is partway to an established baseline. The result shows live meters for sessions, days and keystrokes remaining; the badge unlocks on submissions after enrollment completes.
Signature doesn't verify, cert body was mutated, ledger Merkle root doesn't reproduce, or the archive was manipulated. A valid Solus file cannot become tampered by accident — treat this as a serious finding.
Plain-language answers for reviewers, employers, journalists, and anyone else who's looking at a .solus file for the first time.
unsupported even when the file is cryptographically clean, and the result panel shows live meters for what's still outstanding. This is an absence of evidence, not evidence against anyone. Second, the writer is enrolled but this submission's typing pattern didn't match — either the rhythm shifted (illness, injury, new keyboard) or someone else typed it. Read the Axis B checks for specifics.
Verification runs entirely in your browser using the Web Crypto API and JSZip. No data is sent to X Point's servers or any third party. You can verify this by disconnecting from the internet before dropping a file — it still works.
Students write in Solus free. Teachers verify with one click.