Interpretation & stance
We built Unread to study productive friction because reading tools usually remove the exact struggle that proves comprehension.
Unread takes a hard position: the right AI for reading should not make every text easier. It should make pretending harder.
We built Unread to study productive friction because reading tools usually remove the exact struggle that proves comprehension.
The product refuses the obvious AI-summary answer. It uses AI as resistance: a gatekeeper that says, 'not clear yet.'
Saved rooms, streaming AI feedback, locked sections, presence, reports, public posts, and author analytics all serve one idea.
A stranger should understand it in one minute: paste text, answer the claim, get blocked when you fake it.
Unread is willing to be annoying. That discomfort is the stance, not a side effect.
Open the seeded post: The meeting after nobody read.
Show that only the first section is visible.
Submit a vague answer and let the AI push back.
Submit a precise answer and unlock the next section.
Open the dashboard analytics to show the weak checkpoint.