From research question to Zotero library with PDFs, mostly hands-off
Describe a research question and OpenClaw searches Semantic Scholar and Google Scholar, builds a clean Zotero collection and tries 9 ways to fetch each PDF. It gets about 60-65% automatically.
Done withOpenClawSpotted· by HN user Lowstack· Oct 15, 2026
Result Also runs their to-dos, briefings and meal plans
“I describe a research question and it runs a full pipeline — searches Semantic Scholar + Google Scholar, creates a Zotero collection with clean metadata, then tries to fetch PDFs through 9 different strategies (institutional repos, arXiv, Unpaywall, EZproxy with my university credentials, etc.).”
Spotted: found in the wild. Details are as reported by the source above and have not been independently verified.
What they did
Runs OpenClaw through GitHub Copilot. For a literature review they describe a research question, and it runs a full pipeline: it searches Semantic Scholar and Google Scholar, creates a Zotero collection with clean metadata, then tries 9 different strategies to fetch the PDFs, including institutional repositories, arXiv and Unpaywall.
How it went
It gets about 60-65% of the PDFs automatically. The same setup runs their to-do list through GitHub Issues from Slack, morning and evening briefings, a server-admin agent, and a coach that pushes meal plans to their recipe app. They call it the first AI setup that genuinely changed how they work.
Try it yourself with OpenClaw
Run a literature review on this question: [research question]. Search [Semantic Scholar, Google Scholar] for the most relevant papers from [years], create a Zotero collection called [name] with clean metadata, and try to fetch each PDF from open sources (arXiv, Unpaywall, institutional repositories). Tell me which papers you couldn't get, and give me a one-paragraph summary of the main findings.
A starting prompt written by Agents in the Wild. Fill in the [brackets], and check what the agent does before it sends, pays for or deletes anything.
When a new OpenClaw beta added Telegram message streaming, they asked their agent what else had changed. It read 15+ merged pull requests, compared them to the current version, and gave an upgrade recommendation.
Hospital paperwork needed sorting, some in a language still being learned. An OpenClaw agent scanned 34 pages, flagged missing signatures and drafted an insurance pre-approval for review.
Wanting their son to grow up around books, they hooked OpenClaw to their Goodreads list to scan local used bookstores for titles to buy. They say they've never read as much or enjoyed it as much as now.
An agent box, not a primary laptop: OpenClaw runs on an old MacBook Air, reached through Slack. Its first job, a morning brief with weather, tasks and ideas, came with one broken part: the subreddit scan.
Quotes needed from shower glass suppliers in the area. OpenClaw researched them, sent out the emails, collated the responses and handled the back-and-forth, all in one chat thread.
The current after-school activities end in March, and they'd been told to watch for new ones. Four chat messages had OpenClaw write a skill that checks daily at 10:30am. They still call it unreliable.