Every morning, Vyzhimka goes through your sources and keeps only what you'd actually act on. That's usually 9–11 items — about ten minutes of reading.
What an issue looks like
A sample issue — this is exactly what lands in your inbox.
Anthropic's prompt caching bills repeated context at a tenth of the normal input price after the first call.
Move your agent's static instructions into a cached prefix and compare this week's token bill.
AI · anthropic.com
Postgres asynchronous I/O speeds up cold sequential scans two to three times on NVMe in early benchmarks.
Rerun your slowest analytics query on a test instance before budgeting new hardware.
Engineering · postgresql.org
The EU AI Act's rules for general-purpose models are phasing in; fines scale with global turnover.
Spend an hour checking whether your product's model usage falls under the notification duty.
Regulation · europa.eu
The problem isn't a shortage of news
A feed doesn't end — that's the design. While you scroll, you're not choosing what to read. An algorithm is, and it makes money on your time.
You can't see why you were shown something, and you can't change it. Want something different? Keep scrolling.
So after an hour and forty headlines, nothing's changed — not at work, not in your plans. Tomorrow it starts over.
How it works
You list your sources and sectionsPlain RSS — blogs you already read, tool changelogs, industry press — plus public Telegram channels, added by link. Then your own rules in your own words, like “my team writes Go, skip JVM news.”
Every morning the model throws almost everything outEach item gets one question: what would you do differently, knowing this? No answer — it's out. An empty issue isn't a failure; it means there's nothing worth reading today.
You mark ✓ and −, the filter adaptsTwo buttons under each item. After five days of marks, they start shaping what makes it into your next issues.
How this differs from a smart feed
You can close your feeds — and miss nothingFear of missing out is what keeps you scrolling. Vyzhimka checks all your sources every morning, whether you're looking or not. If it matters for your topics, it comes to you by email.
The selection rules are text you can read and editNot “picked for you by the algorithm” — actual lines you can open and rewrite. If an issue feels off, you can see which rule caused it.
We count how much you didn't readFive items per section, tops. When less is worth reading, the issue gets shorter — not worse.
The archive stays yoursEvery issue is a page on your own blog. Six months in, that's a record of what you found important — not a browsing history.
Also included
Save for laterSaved items collect in their own list — come back whenever you have time.
Deep-dive into a sourceThe “deep dive” button has the model read the whole article and explain what it says, why it matters and who it's for. The result is kept.
Video summaries from a linkPaste a video link and get what it's about and what matters in it. An hour of watching becomes two minutes of reading.
Russian and EnglishSwitch the email and interface language in settings.
new
Public Telegram channels as a source
Add channels with plain links — t.me/name or @name — and their posts go through the same filter as your blogs.
Every morning the model reads the fresh posts and lets through only what changes your decisions.
In the email and on the blog they get their own “Telegram” section.
Public channels only, read through their web preview — no logging into your Telegram account, and we never ask for access to it.
The beta takes up to three channels; the full version has no limit.
This is a beta
It works — the emails go out every morning. I still grant access by hand, so I get to talk to everyone who joins and find out what to fix first.
The trial week includes one deep dive a day. After that, issues keep coming, and we'll discuss full access one-on-one.
Something to improve? Write to me
I read everything. The most useful thing you can tell me is what didn't belong in an issue — that shows where the filter goes wrong faster than any metric.
Six rules that outrank any feature. If something in the product breaks them, we fix the product.
An empty issue is a normal resultIf nothing today changes your decisions, you get an empty issue and an honest number: how many articles we went through. We don't lower the bar to fill an email.
We never show what you missedNo unread counters, no reminders, no “you haven't visited in a while”. There's nothing to miss: the email comes once a day and it ends.
The selection rules are text you editNot “the algorithm picked this” — actual lines. Every item says why it was chosen, and “less like this” writes a new rule right in front of you.
An unmarked item means nothingThree states: useful, miss, nothing. Unread doesn't count as dislike — otherwise the system would learn how busy you are, not what you value.
Full texts are never storedTwo lines and a link to the source. Articles and transcripts stay in memory only until the model has read them.
The archive is yoursEvery issue exports to Markdown with an index. One link in settings and you can take it all and read it without us.
Short answers
How long does it take?
Ten minutes in the morning. Setting up sources takes another twenty, once.
Will my sources work?
If they publish RSS, yes — that's almost every blog, publication and changelog. Public Telegram channels are added with just a link.
What happens to my data?
We keep your marks and settings — the filter doesn't work without them. Full article texts are never kept: just a link and two lines. Unsubscribing deletes everything.
Why an email and not an app?
An email ends. An app with a feed would just be another feed — ours this time.