🔊 Update of CSPaper Review (2025-07-30): user experience enhancements and backend optimizations

[root]

Dear CSPaper Review Users,

We are excited to announce several updates and improvements based on the problems and bugs we captured during the past week. Below you’ll find a summary of the key changes in this release:

User Experience Updates

• Preview Without Sign-in
  To encourage wider use, you can now preview the entire review before signing in. However, please note:
  • You will not be able to save this result to your account.
  • It will make targeted data deletion requests more difficult to process, since the review will not be linked to a specific user.

• Copy Button
  We add a copy button so you can easily copy the content as markdown.

Note: *The “Copy” button is now disabled while the review is still streaming. This prevents incomplete or partial copying and addresses a frequent user request.

Improvements

• Support for Older arXiv Numbers
  The system now tolerates and correctly recognizes older arXiv numbering formats, broadening compatibility with historical preprints.

• PDF Parsing Enhancements

• Title & Author Handling: We enhanced the coverage and robustness when extracting paper meta info from PDF so that you see less empty paper title or authors.

• Page Number Extraction: Parsed PDF content now includes explicit page number information, making it easier to reference specific sections during your review (but not necessarily used in each review).

• Subsection Flexibility for LaTeX: Our Latex processor now tolerate papers written in LaTeX that use \subsection without a preceding \section, ensuring more robust parsing of diverse document structures.

• Author List Display: The displayed author list is now capped at a maximum of two lines for better readability.

• Figure Input Optimization: Figures extracted from PDFs are now normalized in format and resized appropriately before being input into the LLM, potentially leading to improved analysis quality.

Performance & Backend

• Speed and API Optimization
  We have enhanced backend performance and optimized API calls, resulting in faster response times and smoother user experience throughout the tool.

---

We appreciate your feedback and support as we continue to improve. Please let us know if you encounter any issues or have suggestions for further enhancements!

---

Cheers,
The CSPaper Team ️
support@cspaper.org

[root]

Important update:

When users are not signed in, we will hide a part of the review after the streaming finish to drive for sign-in so that we can link a review to a user account. This will enable us to easily delete user review data.