Skip to content

Academic Paper Digest

Example prompt: "Every Friday, search for new papers on large language model safety published in the last week. Summarise the top 5 by relevance, extract the key findings, and send the digest to my email."

How to automate research paper monitoring with GloriaMundo

The Problem

Keeping up with academic research is a losing battle. Hundreds of papers are published every week in any active field, spread across arXiv, conference proceedings, journals, and preprint servers. Researchers and R&D leads know they should be tracking the latest work, but the volume makes it impossible to read everything. Most people rely on social media threads or colleagues forwarding papers, which means they see a biased, incomplete slice of what is being published. Setting aside time for a proper literature scan keeps getting pushed to next week.

How GloriaMundo Solves It

We build a scheduled workflow that searches for new papers in your specified research area every week. A web search step scans academic sources — arXiv, Semantic Scholar, and relevant conference pages — for papers published in the last 7 days matching your topic keywords. A url_extract step fetches the abstract and metadata for any result that the search snippet alone does not cover. An LLM step reads through the abstracts, ranks them by relevance to your specific interests, and writes a concise summary of each — what the paper found, why it matters, and how it relates to your work. The digest is sent to your email as a formatted reading list, and a copy is saved to a Google Doc for your team's reference. Glass Box preview shows you the search results and the drafted summaries before the digest is sent.

Example Workflow Steps

  1. Trigger (scheduled): Runs every Friday afternoon.
  2. Step 1 (web_search): Search for papers published in the last 7 days matching the specified research keywords across arXiv, Semantic Scholar, and relevant academic sources.
  3. Step 2 (url_extract): For top results where the search snippet does not contain the full abstract, fetch the abstract and metadata from the paper page.
  4. Step 3 (llm): Rank papers by relevance, summarise each abstract in 2-3 sentences highlighting key findings and practical implications, and compile into a formatted digest.
  5. Step 4 (integration): Send the digest to the specified email address as a formatted reading list with links to the full papers.
  6. Step 5 (integration): Save the full digest to a Google Doc in a shared research folder.

Integrations Used

  • Gmail — delivers the weekly paper digest directly to your inbox
  • Google Docs — archives each week's digest for team access and long-term reference

Who This Is For

R&D leads and researchers who need to track a fast-moving field — particularly in AI/ML, biotech, and materials science — but cannot justify spending half a day every week on literature review. Also useful for product teams that need to stay aware of academic advances relevant to their technology.

Time & Cost Saved

The discovery and triage part of a weekly literature scan — searching, reading abstracts, filtering for relevance, and writing short notes for colleagues — takes 2-3 hours when done thoroughly. Deep reading of the papers that survive triage still happens manually. Most people either skip the scan or skim whatever appears in their feed. This workflow delivers a curated, summarised reading list every Friday for a few credits per run, replacing the discovery hours rather than the reading hours. Over a quarter, that is roughly 25-40 hours saved, with the added benefit of systematic coverage rather than ad hoc discovery.