Turn new arXiv papers into nightly notes

A daily digest and nightly Claude Code summaries helped one person keep a small, organized notes folder for selected papers.

Done with Claude Code· Intermediate

What they did
They used a Python script tied to a topic file to check arXiv every morning and Google Scholar once a week. The script ignored papers it had already seen and wrote a dated Markdown digest. They reviewed that digest, moved only useful PDFs into topic source folders, and then a nightly script started a Claude Code agent for each new PDF to create a summary note in the matching folder.
How it went
The setup was meant to have each saved PDF summarized by the next digest cycle. The summaries were supposed to include the method, sample, findings, effect sizes, and limitations, but no measured results were reported. A weak prompt could cause the agent to simply restate the abstract.
Tips
Keep a manual review step before summarization so the queue stays manageable. Ask for specific sections and constraints; otherwise the note may be too close to the abstract.

Try it yourself with Claude Code

Help me set up a two-step paper workflow. First, each morning, search arXiv for new papers in [topics] matching [keywords], and once a week check Google Scholar too. Skip anything already seen and write a dated Markdown digest to [digest folder]. Then, after I manually choose which PDFs to keep in [PDF folder], run a nightly agent on each new file and save one Markdown note per paper in [notes folder]. For each note, include the method, sample, main findings, effect sizes if present, and limitations. Before any step that would send, delete, move, or submit files, stop and ask me to confirm. At the end, show me the schedule entries and one example note format.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on chatprd.ai ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.