Loading…
Daily Deck
Each day inveni surfaces a small deck ranked by your personal model. Score the paper to move to the next.Each day inveni surfaces a small deck ranked by your personal model. Tap to give it a like or dislike, it teaches your model. Swipe to view the next paper.
Search
Ask a research question in plain language. inveni expands it, scans the corpus, and reranks the best matches. Takes ~30s.
Profile
Your account, model preferences, and integrations. About inveni & how it works ↗
Personal info
—
Model preferences
Your daily deck of papers is driven by a Deep Learning model. If your interests have significantly shifted, you may retrain it with your new interests input.
Zotero integration
Connect your Zotero library to import papers and save recommendations.
Not connected
Save liked papers to:
library root
Alerts
Choose how you get your daily picks — Email, Telegram, or both. Every digest links straight to arXiv, with AI summaries and figures on inveni.
Telegram
Checking…
Opens @inveni_alerts_bot and binds this chat to your account. Send /today in the bot for the current batch on demand.
Danger zone
Permanently delete your account and all associated data. This cannot be undone.
Model training
A guided setup so inveni knows what to surface tomorrow.
Step 1: define your interests
We need to know what you want to read. Paste a Google Scholar profile URL to bulk-import, or skip and add papers manually on the next step (the search bar there understands author names, paper titles, arXiv URLs, arXiv IDs, and DOIs).
or
History
All your liked papers in one place. Use the navigation buttons to change the week, or open the calendar.
—
Browse by date
Every paper you liked, plus every still-unrated suggestion, from a time frame you pick. Expand a card for its AI summary and figures; rate the unrated ones here.
Admin
System management and diagnostics.
System health
API server—
Supabase—
Host node—
Daily pipeline
Manually trigger the daily recommendation cycle.
Model diagnostics
Run a sanity check (your papers vs random papers).
Users — identity & suggestion quality Full ops dashboard ↗
Onboarding = registration→onboarding funnel (Reg→Prof→GS→Seeds→Onbd→Feed→Active); green = reached. Last seen / Visits·wk / Opens·7d = visit & feed-check frequency. Rated% = engagement. Like% of rated = quality among acted cards. Click a user_id to copy.
| user_id | Last sign-in | Onboarding | Last seen | Visits·wk | Opens·7d | Sug. | Rated | Rated% | Liked | Disliked | Like% rated | 7d | Last fb | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| loading… | ||||||||||||||
Bug reports
In-app "Report a bug" submissions, newest first. Click a thumbnail to open the full screenshot.
loading…