AI Premium pilot
Have a focused AI conversation on an e-ink screen.
Paperyx AI Chat keeps the message box within reach, remembers each thread, and asks for concise answers that are comfortable to read on Kindle and Kobo.
Conversation that carries forward
Each chat is saved as its own thread. Recent messages and a compact hidden memory are sent with the next request so follow-up questions retain their context without making the visible conversation harder to read.
Built around Kindle and Kobo constraints
- A bottom-docked composer remains accessible when the keyboard opens.
- New messages scroll into view automatically, with explicit page-up and page-down controls.
- A small JavaScript enhancement sends messages without a full page refresh and shows an in-message waiting state.
- Normal protected form submission remains available when JavaScript is unavailable.
Usage and cost controls
Administrators choose the shared context, model, response limit, system-wide monthly cap, and default monthly per-user allowance. Administratively licensed pilot users can use Chat without receiving administrator privileges. Each user can have a separate monthly token limit, and usage is recorded from provider-reported token counts. Prompts and answers are not written to application telemetry.
Controlled pilot availability
AI Chat requires an active administratively assigned AI Premium grant or trial, the account AI preference, an administrator-supplied OpenAI connection, and available usage allowance. Public AI Premium checkout and production pricing are not available yet.