Built for agents
across 8M papers.
Live in the agents your team uses.
Hand your agent the whole paper. In one copy.
Every page — and its .mdtwin — is clean, structured markdown: TLDR, full text, figures, genes, proteins, trials, structures, and references. No PDF parsing, no scraping, far fewer tokens.
# CD44: a multifunctional cancer stem cell marker TLDR — CD44 splice variants drive metastasis and mark cancer stem cells; the review maps therapeutic angles. ## Linked entities Genes CD44 · TP53 · EGFR · KDR · MET (+19) Proteins P16070 (CD44) · P04637 (p53) (+22) Trials NCT02753595 · NCT01358903 (+ more) Structures 1UUH · 2I83 (RCSB PDB) ## References (201) · ## Cited by (84)
Every paper. Six sources.
PMC (5M), arXiv (3M), bioRxiv (333K), medRxiv (84K), eLife (20K), F1000 (10K) — one queryable layer for your agent, refreshed continuously.
Decide in 10 seconds, not 10 minutes.
Plain-English TLDR summaries so you can decide in seconds which papers matter. Millions already generated across the corpus, with more every day. Get the gist before you dive in.
From paper to artifact.
GitHub repos, HuggingFace models, datasets on Zenodo, crystals on COD, structures on PDB — all linked directly to the paper.
Know which papers to trust.
OpenReview peer reviews, retraction status, FWCI impact scores, citation context. The trust signals before you cite — or before a clinical decision.
Watch the walkthroughs.
Conference talks, Yannic Kilcher, Two Minute Papers. Instantly watch explanations right alongside the math.
Every gene, drug, and clinical trial.
Papers linked to the proteins, chemicals, diseases, clinical trials, and PDB structures they mention — pulled from PubTator, UniProt, gene2pubmed, ClinicalTrials.gov, and RCSB PDB.
Every gene. Every paper.
1.58 million papers linked to their genes via gene2pubmed — the NCBI authority. BRCA1, TP53, EGFR, KRAS, and 60K+ human genes resolve directly to the source publications.
Proteins to publications.
477K UniProt accessions linked to their source papers. P04637 (p53), P38398 (BRCA1), Q9Y6K9 — every Swiss-Prot entry traceable to the publications that characterized it.
Trace any trial to its papers.
132,634 papers linked to ClinicalTrials.gov NCT IDs — every phase, every status. Find the foundational science behind any registered trial, or every paper a single NCT cites.
From sequence to structure.
108,980 papers linked to RCSB PDB structures. Crystallographic data, cryo-EM solutions, NMR ensembles — every 3D structure resolves to the publication that solved it.
Map the entire literature.
25,380 semantic clusters across 8.24M labeled papers — UMAP + HDBSCAN over the full corpus, topic-labeled via OpenAlex. Find a paper’s neighbors instantly, no keyword guessing.
Want early access to new features?
No spam. Just big updates.