Paper Digest indexes 1,200 ICLR papers

Not a launch in the usual sense, but a high-signal curation product for a conference that’s already too big to track manually.

• –The roughly 1,200 listings amount to about 22% of ICLR 2026’s 5,300+ accepted papers, which is a reminder that public artifacts are still the exception, not the default
• –Because the index is auto-extracted from submissions, it is fast and scalable, but some links will lag release timing or miss code that stays private until the conference opens
• –For developers, the real value is filtration: it turns a giant accepted-papers dump into a shortlist of repos, datasets, and demos you can actually inspect and reuse
• –Paper Digest is effectively positioning itself as research infrastructure, combining conference indexing, summaries, and workflow tooling around the literature