OpenAlex

@OpenAlex search for recent papers on transformer architectures in natural language processing and show me the top 5 by citation count

Here are the top 5 papers on transformer architectures in NLP by citation count: 1. "Attention Is All You Need" (Vaswani et al., 2017) – 89,432 citations 2. "BERT: Pre-training of Deep Bidirectional Transformers" (Devlin et al., 2019) – 67,821 citations 3. "Language Models are Few-Shot Learners" (Brown et al., 2020) – 34,567 citations 4. "Improving Language Understanding by Generative Pre-Training" (Radford et al., 2018) – 28,943 citations 5. "XLNet: Generalized Autoregressive Pretraining" (Yang et al., 2019) – 19,234 citations Each entry includes DOI, abstract, author affiliations, and open-access status. This is an illustrative example; actual results depend on OpenAlex's current index.

This example demonstrates OpenAlex's core strength: comprehensive academic metadata search across 250M+ works. The autocomplete and metadata retrieval tools work together to resolve queries and return structured bibliographic data. No special auth scopes required beyond a valid API key, but complex queries count against your daily rate limit.

@OpenAlex create a collection called 'Q1 Literature Review' and add these five papers to it so I can track them for my research project

Collection 'Q1 Literature Review' created successfully (ID: col_abc123xyz). All 5 works have been added: - W2964141474: "Attention Is All You Need" - W2807374136: "BERT: Pre-training..." - W3037662171: "Language Models are Few-Shot..." - W2809633644: "Improving Language Understanding..." - W2963436189: "XLNet: Generalized Autoregressive..." You can retrieve this collection anytime using its ID. This is an illustrative example showing how the create and bulk-add tools coordinate to organize research materials.

This showcases OpenAlex's collection management for organizing research. Collections are private to your account and persist across sessions. Important: collections are single-entity-type only (works, authors, etc. cannot mix), and deleting a collection does not affect the underlying OpenAlex entities. Useful for literature reviews, grant proposals, or collaborative research tracking.

@OpenAlex check my rate limit status, then download the full-text PDF for the Attention Is All You Need paper if I have budget remaining

Rate limit check: You have 847 queries remaining today (resets in 14 hours). Prepaid balance: $12.40. PDF download cost: $0.01. Downloading PDF for W2964141474 ("Attention Is All You Need")... Complete. The 15-page PDF has been retrieved from OpenAlex's cache. Note: This is an illustrative example. Actual downloads depend on whether OpenAlex has cached content for the work (check has_content flags first), and you must comply with the original publisher's copyright terms even though OpenAlex hosts the file.

This example highlights two critical tools: rate-limit inspection before expensive operations, and content download for works with cached PDFs or GROBID XML. Each download costs $0.01 from your prepaid balance, so always check budget first. Not all works have downloadable content—verify has_content flags in the metadata before attempting. Downloaded files remain subject to original copyright restrictions.

When OpenAlex beats manual citation chasing for small research teams

A three-person academic lab writing an NIH R01 needs 40-60 recent papers on CRISPR delivery mechanisms, filtered by citation count and open-access status. OpenAlex's autocomplete and metadata tools let you build that list in under an hour without touching PubMed's clunky interface or paying Scopus fees. The collection tools keep your shortlist in one place while you draft the background section. The catch: if you need full-text analysis of paywalled papers, OpenAlex only caches open-access PDFs—you'll still hit institutional login walls for half your sources. For teams writing one or two grants a year, the API key is free and the workflow is faster than spreadsheet hell.

Track competitor publications without a $15k Scopus contract

A six-person biotech startup monitors 12 competitor labs by author and institution, pulling new publications weekly to spot IP risks or partnership signals. OpenAlex's author and institution metadata tools make this a 20-minute Slack standup ritual instead of a half-day manual search. You create one collection per competitor, add entities as they publish, and export the list to your CTO. The rate limit status tool matters here—if you're downloading dozens of PDFs for patent prior-art searches, the $0.01-per-download fee adds up fast. For monitoring and triage, OpenAlex is the right call. For deep patent work, you still need a real IP database.

When OpenAlex works for episode prep but not live fact-checking

A two-person science podcast team preps episodes on topics like microplastics or fusion energy by pulling 15-20 landmark papers, checking author affiliations, and verifying funding sources. OpenAlex's metadata retrieval and funder tools give you that context in minutes, and the autocomplete helps when a guest mentions a paper title mid-interview. The workflow breaks if you need real-time lookups during recording—API calls take 2-4 seconds, and the interface isn't built for live search. Use OpenAlex in the research phase, then screenshot or print the key facts. For a weekly show, the free tier covers your volume and the metadata depth beats Google Scholar.