Truly Typed – A writing app for the AI era
Keystroke dynamics to catch transcription is a genuinely clever anti-cheat mechanism.

Tackles real frustration (AI slop in comments), but typing signals are easily spoofed.
HN users, content creators, anyone fighting AI-generated spam in comment sections
Originality.ai · Turnitin · GPTZero
The tool today measures high level signals (typing consistency, paste %, edit rate ,etc.) to give a score. After writing your message you save an attestation link and include it in the post/comment. This basic approach is deeply flawed, but hopefully better than nothing. I'm curious what current state of the art is for _guaranteeing_ that something is human typed. Perhaps some sort of hardware integration?
I'm working on a browser extension, and a quick way to integrate attestations with comment sections so users don't need to include the link copy/pasted. Would love any and all feedback.
Keystroke dynamics to catch transcription is a genuinely clever anti-cheat mechanism.
Keystroke + audio forensics for authorship proof, but keystroke spoofing is researched and brittle.
AI adoption dashboard for GitHub, but detection accuracy fully depends on commit message patterns.
Typing stress detector for macOS, but wellness apps already do this better.
Proves AI usage with a verifiable PDF receipt instead of probabilistic detection guesswork.
Cryptographic proof of human writing process beats broken AI text detectors.