01 Independent project · 2026
Academious
A personal research feed. Tell it what you work on, and each morning it brings you the handful of new papers worth your attention.
The new home page: set up your interests once, and new papers come to you instead of you searching for them.
Search is still there when you need it: an everyday phrase returns the most relevant papers first, clearly labelled when they are free to read.
How it works
From five open sources to the handful of new papers worth your morning.
Collect
Gathers new papers every day from five open sources.
PythonClean up
Merges duplicates and flags retracted or open-access papers.
PostgreSQLMatch
Compares each new paper with your fields and topics, by meaning as well as keywords.
Machine learningLearn
Your feed sharpens as you open, save or skip papers.
React websiteAccounts, a personal feed and saved collections are live. Personalisation is opt-in, and deleting your account deletes everything.
01 Problem
Why this exists
I wanted to keep up with research related to my interests, but searching Google Scholar buried me in papers that were technically relevant and not actually useful. I didn't need a bigger list of results. I needed a calmer way to find the few papers worth my time.
Academious is live at academious.org and holds over half a million papers from arXiv, bioRxiv, medRxiv, Europe PMC and OpenAlex. The latest version turns it from a search box into a feed: you pick your fields and topics in about a minute, and new papers that match are waiting for you each day.
02 My part
What I built
Collected papers automatically every day from five open sources, including arXiv, bioRxiv and Europe PMC.
Cleaned and merged them into one collection, removing duplicates and flagging retracted papers using the Retraction Watch database.
Built a search that understands the meaning of your question, not just the exact words.
Added accounts and a personal feed that matches new papers to your interests and sharpens as you open, save or dismiss them.
Added saved collections, an interests page and clear open-access labels, and launched the redesigned website.
03 Approach
How I approached it
Good data first
A feed is only as good as the papers behind it, so I treated collecting and cleaning them as the core of the product.
Bring papers to people
Most people don't want to search every day. A short daily list that learns from you beats an empty search box.
Your data, your call
Personalisation is opt-in, and deleting your account deletes everything it learned.
04 Result
Where it is now
A live research feed with accounts, daily recommendations that learn from what you open and save, saved collections, and search across one combined collection of over half a million papers.
It only covers papers from its five sources, not all published science. The feed needs a little use before its suggestions get sharp, and I haven't published quality scores yet because I want people to judge the results first.
Next: Measure how useful the daily suggestions really are with early users, and publish quality results once people have judged them.