ApartureBringing the arXiv into focus
One briefing a day replaces the listings. Every citation in it is checked against the papers it was given.
One briefing a day replaces the listings. Every citation in it is checked against the papers it was given.
Aparture is a local web app that reads the day's arXiv so you can stop skimming it. You describe your research in plain English and pick the arXiv categories to watch. Each run scores the new preprints in those categories against your description and reads the strongest ones in full. It then writes a briefing that groups the papers into themes and says why each one matters for your work. Every citation in the briefing is checked against the papers it was actually given before anything renders.
It's designed to be used over weeks rather than once. You star papers you liked, dismiss ones that missed the mark, and comment where you have specific reactions. When you ask for suggestions, Aparture turns those signals into proposed edits to your profile, and you accept or reject each one.
You describe your research the way you would to a colleague, not as a keyword list. Every stage reads that description, from the first filter to the briefing.
Each run produces one piece of writing that argues across the day's papers. It opens with an editorial lead and tells you which papers to open first.
Star, dismiss, or comment on what you see. The Suggest Improvements flow reads that feedback and proposes specific edits to your profile, each with the feedback that motivated it. You accept or reject each edit.
Aparture works with Anthropic, OpenAI, and Google models, and each stage can use a different one: a cheap model for filtering, a stronger one for reading papers. Your keys stay in .env.local on your machine.
Node, the repo, optional Playwright. About 5 minutes.
STEP 02Pick a provider, create a key, paste into .env.local.
Dry run (free) and Minimal API Test (~$0.20–$1 on paid tiers).
A narrated 10-minute walk through your first real briefing, end to end.
Already have Aparture running? The Guide covers daily use, from running and reading briefings to giving feedback and refining your profile. Under the Hood explains how the pipeline and briefings work. Reference is where to look up environment variables, prompt files, and troubleshooting symptoms.