Introducing daily-academic.com a smarter way to stay updated with scientific research
Hi everyone,
I wanted to share a project I’ve been working on: daily-academic.com a platform designed to help researchers and bioinformaticians discover relevant scientific papers more efficiently.
What it does:
Mobile friendly web page
Why I built it: Keeping up with the literature is overwhelming, especially with the volume of new publications every day. I wanted something lightweight, customizable, and focused without the noise.
Planned features:
Integration with GEO / SRA datasets
I’d really appreciate feedback from the community especially:
If you're interested, feel free to check it out and let me know what you think!
Link: https://www.daily-academic.com
Thanks
I actually gave this a quick try and it’s surprisingly useful. The keyword based filtering works better than I expected, and it’s nice to see something lightweight instead of the usual overly complicated tools.
What I liked most is that it doesn’t try to do everything it just helps you find relevant papers faster. Curious to see how it evolves.
Some potential improvements could include allowing users to integrate their Google Scholar profile, enabling the platform to recommend relevant literature based on their publication history particularly emphasizing papers where they are first or last author.
Google Scholar profile will be a nice touch. Althought google is not very happy giving data!
Congrats and thank you for setting up such useful and clean tool. I gave it a try and it is indeed quite easy to grasp.
For the cards :
- Add the logo of each journal in the card
- I don't find it helpful to display the first 20/30 words of the abstract
- Could you get the nationality of the lab/last authors (from ORCID or references) and display a small flag on the card
Thank you for the thoughtful feedback! I really appreciate you taking the time to share this!
- Add the logo of each journal in the card
I agree this would be a great visual addition. However, I’m currently facing some challenges sourcing journal logos, and I’m also looking into the legal side of using them. For example, when I recently published a paper in NAR, they specifically asked us to remove all journal logos from figures, so I want to make sure I stay compliant.
- Abstract preview
That’s a fair point. Would you prefer seeing the full abstract directly on the card? I’m planning to include the complete abstract within the card in the next deployment.
- Could you get the nationality of the lab/last authors (from ORCID or references) and display a small flag on the card
I really like this idea as well. I’ll explore whether I can reliably retrieve nationality information (e.g., via ORCID or affiliations) and display it as a small flag on the card.
Thanks again, this kind of feedback is super helpful for improving the tool!
I guess it would also be a legal nightmare to do an AI summary of 10 sentences for each paper.
As an other comment mentioned, it would be helpful to export the bookmarks to get DOI in an Excel sheet (so we can do this AI summary externally).
Any reason why only the like is available for preprint ? Would be nice to be able to bookmark them.
Would be cool to be able to move the bookmarked papers between folder (now one needs to delete the bookmark, search for the paper again and bookmark to another category).
Would it be possible to pin author names or labs (by name or Orcid ID) to get updated of their next publications ? Maybe another section in the menu.
I got another idea related to the previous one, in "manage account" why not allow the user to register its Orcid profile. Have a section "Your papers" and see how many people upvote, comments or bookmarks your paper.
And on top of this, you could even refine your algorithm by : "Among the users that bookmarked your paper, these are the other papers that they have bookmarked" (more or less how spotify is working if I recall).
Depending on how many people you get involved in this, you could also do a scientific paper network graph based on the bookmarked or like system to find potential labs to collaborate with etc...
Noted!
Orcid and google scholar profiles will be definitely added into the system.
The recommendation algorithm will be tuned over time to find patterns based on other users similar interests.
Log in to answer this question.
People may definitely pay for a tool that seems to be useful based on the comments below but it would be best to indicate what the "Subscription" is going to cost. Would there be a "free" tier always available?
Yes sure. The plan is to introduce subscriptions to help cover server and cloud hosting costs. At this point, the infrastructure is still relatively inexpensive, so I’m covering it myself during early access.
I guess you are using llms to create a recommendation system. Do you use public APIs eg openAI?
We’re actually using in house models for the recommendation system rather than relying on public APIs and we try to avoid the cost. At the moment, we’re also preparing a bioRxiv paper (with the goal of later submitting it to a peer-reviewed journal) that will describe the methodology and architecture in detail. I will post it here when we submit it!
The screenshot looks impressive. Can you elaborate how your tool differs from existing ones?
Since about a year, most Swedish universities offer a campus license for Keenious, so that is what I currently use. At least a few German institutions that collaborators work for signed up for Avidnote. GrantflowAI already went out of business again and open-sourced their software. So it is evidently a highly competitive market and there must have been a feature you were missing elsewhere, if you decided to build on your own?
This is in no way meant to discourage you, but as the notorious reviewer 2, I would like to better understand the novelty of tool :-)
PS: I remember a talk from somebody at EuropePMC delivered about a year ago who said that they are currently building a public vector database with embeddings of all papers indexed by EuropePMC. I never bothered to check if that project went public since, but at that talk he hinted that the intention was to offer this as an open endpoint similar to the traditional search. If its live now, that would probably help you a lot in case you haven't already integrated that?
Thanks a lot for the detailed comment this is exactly the kind of “reviewer 2” perspective thats actually helpful!
You are absolutely right that this is becoming a very competitive space, and tools like Keenious are strong products. Our motivation to build something came precisely from gaps we felt in our daily workflow.
For example, Keenious is primarily a chat based interface. It works very well when you already know what you are looking for (e.g., asking something like “what is the best algorithm to detect TSS with high recall?”). It is excellent for targeted queries.
What we are trying to build with our tool (Daily Academic) is a bit different in philosophy. The focus is less on searching for answers and more on discovering what is new and staying continuously updated. In practice, that means:
So the distinction is closer to:
Overall, our goal is NOT to compete head on as a generic LLM powered assistant, but to focus on a more passive discovery + domain specific intelligence layer that fits naturally into the researcher daily routine.
Also, to clarify this is not something that we are trying to turn into a heavily commercial product. Down the line we might add a small subscription just to cover server costs, since running LLMs is quite hardware intensive. The idea would be to keep it much lower than tools like Keenious, which are usually around 20 Euros to get the most out of it.
How often do you udpate the new publications of each journal ?
I am looking for the paper "Whole-body molecular and cellular mapping of the laboratory mouse" publish in Cell on March 27th, but going in the "Source Journal" section (https://app.daily-academic.com/JournalArticles?journalId=25166&journalTitle=Cell), the last updated paper was on March 21st.
It depends on the journal and how often it syncs with public databases. In some cases, there can be delays or gaps if the journal doesn’t update consistently with sources like PubMed, so newer publications might not appear immediately.
The system updates daily.
I have opened a new discussion on GitHub about this thanks for pointing it out! It will be updated by the end of May.
https://github.com/dailyacademic/support/discussions/1