← All work

LTUNE DAILYMy product · 2026 · AI newspaper for Kindle

Each night AI lays out a newspaper for my Kindle.

Task: Mornings started with the phone: reels, messages and other people's urgent stuff before I was even awakeSolution: built a system that reads my channels overnight, strips ads and lays out a paper for KindleResult: since mid-November an issue arrives every morning: newspaper first, everything else later

By 6:30 an issue is on my e-reader: news from my channels, ad-free, a project briefing and archive photos from this day. I grab the Kindle, not the phone.

Built the first working version, from layout to sending from the server, with AI agents in one evening. Then tuned it issue by issue, reading each one on the Kindle.

LTUNE DAILY

One morning: phone or newspaper

What you see in the first minutes after waking up
Before · phone
  • Pick up the phone and get reels, notifications, messages and other people's urgent stuff.
  • The feed never ends: news, ads and retellings of other people's posts all mixed together.
  • Not even out of bed, and your head is already full of someone else's day.
✦ Now · LTUNE DAILY

At 6:30 the issue is already on the Kindle: a cover, the story of the day, AI and design pieces, a briefing on my work and photos from this day in my archive. Ads and retellings were filtered out overnight. Four pages, and the newspaper has a last page.

Issue by 6:30No ads or retellingsThere is a last page

Why it's harder than it looks

01Filtering before the editor

Calling a big model on every post is expensive. A fast sorter, Jev, sits before the editor: it writes nothing, only answers closed questions and drops a fifth to a third.

02A newspaper, not a blog post

Models love "in this issue we will tell you". Strict rules fixed it: the paper never talks about itself, every paragraph has a hard fact, headlines explain the event.

03Kindle broke the layout

Send to Kindle turned a vector PDF into a hundred pages of mush. Now pages render at 2160×2880, photos get a newspaper halftone, the issue ships as images in EPUB.

What's inside

An issue on Kindle Scribe
An issue on Kindle Scribe
Cover of issue one
Cover of issue one
Page III, industry
Page III, industry
Collectors

Telegram via web preview without API, media, GitHub releases, industry news.

Briefing from my knowledge base

Tasks, calendar, project statuses and a week of work session summaries.

Jev sorter

Is it an ad, is there substance, which section, how important.

Claude as editor

Writes the issue by the style guide, only big news makes the front.

Magazine-grade layout

New Yorker and the old Afisha as references, Cyrillic-ready fonts only.

Photos of the day

Shots from my archive taken on this day in past years.

How the pipeline works

Same thing every night: the server gathers material, AI models filter and write, Chrome lays out the pages, and the Mac prepares photos ahead of time.

01Collection and circuit breaker

A day brings 60 to 150 items. At most 150 go into work. If a collector brings three times that, the issue stops before the sorter: a failure, not a harvest.

02Filtering and editing

Jev answers closed questions, plus two for memes: funny without context, readable on e-ink. Claude Sonnet gets the clean selection with the style guide and writes every page.

03Layout and delivery

Jinja and headless Chrome build the pages, halftone matches the screen's pixels. At 6:30 a server timer sends the EPUB to Kindle. The Mac picks photos locally with Vision.

How an issue is built
How an issue is built
The sorter's questions
The sorter's questions

Shall we discuss your task?

Half an hour of conversation: I'll look at the task and tell you honestly whether AI is needed here and where to start. No presentations and no 40-slide proposals.