The reading companion · No. 01
Save any link. Get the essence in ten seconds.
Paste a URL. We read it, we distill it, we hand you back a hundred and fifty words, five clean bullet points, and one line worth keeping. That is all. You were going to skim it anyway.
A sample, unedited
Why the slow web is coming back
The last decade taught us to scroll. The next one will teach us to slow down. In this essay, the author argues that the infinite feed has quietly emptied the web of its middle ground — the essay, the journal, the considered note — and that readers are starting to push back. RSS is up. Newsletters are up. Even handwritten letters have found a second life. What remains uncertain is whether the institutions built around attention can survive a readership that refuses to keep scrolling.
- The feed optimizes for novelty, not for depth. Most users now feel both exhausted and uninformed.
- RSS subscriptions in 2025 surpassed their 2009 peak, quietly, without a single conference talk about it.
- Newsletter retention beats social retention by a factor of three, but only for writing that takes longer than twenty minutes to read.
- The slow web does not mean anti-technology. It means anti-interruption.
- The next platform war may be about what to remove, not what to add.
"We spent a decade teaching people how to skim. We now have a generation of readers who never learned to finish."
Example · The actual summary was produced in 9.2 seconds
How it works, in three lines
You paste the link.
Anything. A longform essay, a research paper, a New Yorker profile, a ten-thousand-word blog post you promised yourself you would get to.
We do the reading.
We fetch the page. We strip the ads, the cookie banners, the sidebars, the six unrelated newsletters it tried to sell you. What remains is the text the author actually wrote.
You read the essence.
A hundred and fifty words, five bullets, one quote. Structured. Exportable. Searchable later when you actually need it. You decide whether the original is worth the longer read.
We did not build this to replace reading. We built it to protect the reading that matters from the reading that doesn't. From a letter to our first twenty users
Machine-readable
JSON, Markdown, plain text.
Everything we return can be exported the moment you receive it. Pipe it into Obsidian. Drop it in a Google Doc. Feed it to another model. Your notes, your format.
Folders and tags
A small library, kept on purpose.
Save the summaries worth keeping. Tag them. Search them. We don't try to build a social graph. We try to build a better commonplace book.
Nothing loud
No push, no red dot.
aisummary doesn't ping you. It doesn't grow your streak. It opens when you open it and closes when you close it. A tool, not a habit.
Private by default
Your reading list is your own.
We don't sell your URL history. We don't train on what you save. The summaries you keep are visible to you and to no one else. We wrote it into the Terms on purpose.
Two plans. That's it.
Free for readers. Eight dollars a month for readers with too many tabs.
Free gets you ten summaries a day. Pro gets you as many as you need, folders, tags, the API. No enterprise tier. No sales calls. No contracts to sign.
A small note from Portland
There are three of us, and we made this for ourselves first.
We had a folder called read later that became read never. We built this to fix that. If it helps you too, we'll be glad. If it doesn't, please tell us why.