Nimbus Feed reduces hundreds of RSS articles to a 15-minute daily digest. AI filters the noise. You read what matters.
You open your reader Monday morning. 1,847 unread. You skim the first 30 headlines — half of them are the same story from different sources. You star two articles you'll never go back to. You mass-mark-read the rest and feel vaguely guilty about it.
Tomorrow it's 1,847 again.
You've tried rules. You've tried folders. You've tried “just unsubscribing from a few.” But the problem isn't how many feeds you follow — it's that your reader shows you everything and helps you with nothing.
That's not a reading problem. That's a tool problem.
Every morning, Nimbus Feed reads your feeds and generates a Morning Brew-style briefing — grouped by topic, ranked by relevance. You write the “Recipe” (a plain-language prompt for each feed group), and the AI follows it. Pick your model. Set your token budget. If the budget runs out, it falls back to raw excerpts — no silent failures.
One digest. 15 minutes. 200 feeds covered. Newsletters included — ~90% have RSS feeds, and Nimbus finds them automatically.
When 12 sources cover the same Apple earnings story, Nimbus Feed groups them under one header — Google News-style. The duplicates collapse. The noise drops. Mute rules work in plain English: “mute anything about cryptocurrency unless it mentions regulation.” Every muted article shows the AI’s reasoning and an “Allow this” button. Nothing is deleted. Everything is auditable.
Two-stage dedup catches “Apple earnings” vs. “Google earnings” — no false positives.
Highlights sync to your Obsidian vault. Append-only — Nimbus never overwrites your edits. Articles tracked by YAML frontmatter IDs, not filenames. Rename your files. Reorganize your vault. Sync keeps working.
Jinja2 templates for full control over the output format.
Split pane. Infinite scroll. 847 unread. You spend 45 minutes deciding what to read before you read anything. By the time you're done sorting, you're out of time for reading.
Inbox shows new items with time estimates — “~30 min of reading today.” Tap s to send to Queue, m to mute, j/k to move through the list. Queue holds what you chose to read. Nothing else.
47 minutes scanning → 8 minutes triaging. The rest is reading.
Most “AI features” want to show you more. More recommendations. More suggestions. More content you didn't ask for.
Nimbus Feed's AI does the opposite. It reads everything so you don't have to, then tells you what to skip. That's it.
You set the rules in plain English. The AI follows them. When it filters something you wanted, you tap “Allow this” and it learns. The mute audit trail shows every decision the AI made and why.
Your feed. Your rules. Transparent AI that explains itself.
You've organized your Obsidian vault carefully. You've renamed files, built backlinks, structured folders your way. The last thing you need is a sync tool that overwrites your work or breaks when you rename a note.
Nimbus Feed's PKM sync is append-only. It adds highlights and articles to your vault. It never modifies what's already there. Articles are tracked by YAML frontmatter IDs — not filenames. Rename, reorganize, restructure. Sync keeps working.
Customize the output with Jinja2 templates. Control what lands in your vault and how it's formatted. Your vault is the source of truth. Our servers are a pipeline, not a destination.
Append-only. ID-tracked. Your edits are sacred.
No free tier with ads. No enterprise tier at $200+/mo. No “contact sales” button hiding the real price.
One plan. Every feature. $8/mo or $72/year (save 25%).
Our AI costs us $2.07/mo per user at 500 articles/day. We charge $8. That's a sustainable business, not a VC growth hack.
vs. enterprise readers
Real AI starts at $200+/mo. Consumer tier gets keyword search and ads.
$8/mo. Full AI. No ads.
vs. read-later apps
Beautiful for 10 articles/week. Freezes at 150/day. PKM sync breaks on file rename.
Built for volume. Sync tracks by ID, not filename.
vs. rule-based readers
Powerful if you maintain 47 Boolean filters on weekends.
Plain English mute rules. The AI handles the rest.
Detailed comparisons: vs. Feedly · vs. Readwise Reader · vs. Inoreader
You've seen platforms disappear. RSS readers acquired and sunset. Two-week notices to export your data. We get it.
Nimbus Feed charges $8/mo because it costs us $2.07/mo per user to run the AI. That's a 74% gross margin — a real business, not a growth-at-all-costs bet. Your feeds live in standard OPML. Your highlights live in your Obsidian vault. If we disappeared tomorrow, you'd lose nothing.
We'd rather you stay because the product is good, not because leaving is hard.
Import your OPML. Set up a Recipe. Wake up to a digest instead of 1,847 unread.
No credit card required. Bring your feeds. Take them anytime.