Q: Afforai /Logically License Tier 3 user. Questions:
As a pharmacy wanting reliable and up-to-date academic access, we need to ensure the tool meets our research and compliance standards.
1 Which Open Access databases and sources do you index?
2 Do you include PubMed, Europe PMC, DOAJ, medRxiv, BioRxiv, and other major repositories?
3 Are peer-reviewed journal articles, clinical trial reports, and systematic reviews included?
4 Do you index or provide summaries of paywalled or subscription-only research (e.g., Elsevier, Springer, Wiley, NEJM, JAMA)?
5 How do you handle articles that are not Open Access—can I access abstracts, or just see citations?
6 How frequently is your database updated to include recent publications?
7 Do you have the capacity to deliver articles published in the last few weeks or months?
8 Will an future update allow us to upload our own PDF's/scientific papers?
Joe_Wonders
Edited May 31, 2025A: Hi MenScript, happy to answer! We are developing Wonders to be suitable for PhD level workflows—with control & transparency being our top priorities.
1. We index pretty much all available OpenSource indexes. We also lean on semantic indexes, so our coverage is pretty broad.
2. Yes, all those should be included.
3. Yes, all these formats should be available, but you'll need to test out specific sources.
4. We do index paywalled articles, they will surface in searches unless you exclude them using filters.
5. You can access abstracts of all articles. We also use those + metadata for AI processing.
6. Our databases are updated pretty much in realtime.
7. Yes, you can the Year filter to narrow in on the last year, and then sort by Newest (this may reduce relevancy of our ranking system, so you might need to place a little more work into your search keywords)
8. Yes. PDF upload is our #1 request and it's on the roadmap. Welcome to support the idea here: roadmap.readwonders.com
Hope this helps answer your questions! Feel free to shoot us a message on support@readwonders.com or chat with us from our site if you need more clarifications.
Joe
MenScriptPLUSSo I assume Wonders AI uses full text for open access papers, but for paywalled articles it relies only on the abstract and citation info. So answers from paywalled sources are limited since it only “knows” what’s reported in public fields since it cannot actually assess the full methodology, sample size details, exact results, or limitations, unless these are summarized in the abstract?
Verified purchaser
That's absolutely correct.
We're hoping to overcome this with PDF upload, institutional logins, subscription linking, and ability to purchase papers. But I can't promise exact timelines on this, as some of it is dependent on our growth and publishers willingness to collaborate.
MenScriptPLUSPurchasing papers isn’t viable; uploading PDFs is essential. I may have another option, which I’ll email privately. I’ll test your service to see how it performs and if it is immune to hallucinations. PS: Please prioritize a “memory” function, as repeated mistakes are a major issue with academic AI tools and I've never encountered one that didn't require repeated adjustments.
MenScriptPLUSJust to clarify, I’m not criticizing the tool—I just want to make sure I understand how it works and am aware of its limitations. Final question: If I point out that an answer provided is incorrect or incomplete, is Wonders AI able to adjust future responses based on this feedback, or does it treat each query independently? I.e does it have "memory"?
Verified purchaser
Understood. I appreciate you asking these questions—it clearly shows you know what you're doing, and helps us anticipate needs for the most rigorous use-cases.
Each query is independent at the moment.
This video might help explain the workflow:
https://help.readwonders.com/en/article/video-getting-started-with-wonders-t37rfa/