AskChem indexes 2.4M claims from 147K papers

AskChem has indexed 2.4 million source-grounded claims from 147,000 chemistry papers, exposing them to AI agents through REST, SDK, and MCP interfaces, according to the system’s July 30 paper. On AskChem-Bench, grounding GPT-5.5 in the index produced resolvable DOIs for 100% of citations, versus 88.3% without retrieval. For teams building the literature-synthesis layer above domain-specific scientific models, the useful fact is not another chatbot score: AskChem offers a live, provenance-carrying corpus that can make every generated chemistry claim traceable to a paper and evidence locator.