01
Avalanche-assisted upconversion reaches the single-nanocrystal limit
Why it matters Directly connects single-particle dynamics with nonlinear gain.
99FIT
Local-first research intelligence
Your research feed,
distilled.
Turn hundreds of unread Inoreader items into a focused daily brief, ranked against your CV and the work you actually care about.
01 Read-only Inoreader
02 Encrypted local library
03 Open source
YOUR DAILY RESEARCH BRIEF
3 papers worth your time
02
Interface design unlocks a new route to coherent energy transfer
Idea spark A transferable control strategy across material platforms.
86FIT
EDITOR'S NOTE
The strongest theme today: controlling heterogeneity instead of averaging it away.
01 / THE SIGNAL
“Not another feed.
A decision about what to read.”
Daily literature volume keeps rising. Your attention does not.
PaperPulse stops at the useful papers—even when that means returning fewer than your chosen maximum.
02 / HOW IT WORKS
From unread to understood.
A practical research pipeline, built around your interests rather than generic popularity.
-
01
Collect the unread
PaperPulse reads new, unread items across all your Inoreader subscriptions and folders.
-
02
Build your lens
Upload a PDF or DOCX CV, then review and edit the research profile it creates.
-
03
Rank with context
A two-stage AI pipeline scores fit, frontier value, and credible cross-field inspiration.
-
04
Learn from feedback
Relevant, inspiring, known, saved, and not-useful signals improve future briefs locally.
03 / THE BRIEF
Enough context to decide.
Not enough to drown.
Every recommendation earns its place with a concise, structured case—not just a similarity score.
01
NATURE PHOTONICS · PHOTONICS
Avalanche-assisted upconversion reaches the single-nanocrystal limit
99FIT
CORE FINDING
What changed, how it was shown, and where the novelty sits.
WHY IT MATTERS
One direct reason this article deserves your attention today.
YOUR CONNECTION
The concrete link to your methods, systems, or active questions.
IDEA SPARK SPECULATIVE
A testable direction, clearly marked when it crosses into inference.
01
Your chosen N is a ceiling
Ask for 20. Get 4 when only 4 are worth showing.
02
Control the exploration
Switch between strict, balanced, and exploratory discovery.
03
Keep the original one click away
Every brief links directly to the source article.
04 / YOUR CONTROL
A radar you can tune.
Choose how far the radar reaches.
Stay close to your current work or invite carefully labeled cross-field ideas.
Boost, lower, or exclude.
Search every past brief.
Your refresh history stays available in an encrypted local archive.
“Show me the best 20” never means “invent a reason to show 20.”
05 / PRIVACY
Your library stays yours.
PaperPulse is designed to run on your own computer. It does not turn your research history into a hosted PaperPulse account.
- LOCALEncrypted at rest
CV files, OAuth tokens, article cache, recommendations, embeddings, and feedback are encrypted locally.
- READ ONLYInoreader stays untouched
No automatic read, star, save, or subscription writeback.
- TRANSPARENTYou choose the AI provider
Your research profile and article text are sent only to the OpenAI API you configure for ranking and analysis.
- OPENInspect every line
MIT licensed, public on GitHub, and built without analytics or tracking.
06 / START
Make tomorrow's feed
smaller and better.
Requires Python 3.11+, Node.js 22.13+, Inoreader Pro, and an OpenAI API key.
Read the complete setup guidePowerShell
# 1. Download PaperPulse
git clone https://github.com/Jiaye1998/PaperPulse.git
cd PaperPulse
# 2. Install and configure
.\setup.ps1
# Add your keys to .env
# 3. Start your research radar
.\run.ps1READ LESS. DISCOVER MORE.