AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis
AskChem indexes chemistry papers as source-grounded claims instead of ranked documents.
The system converts papers into atomic, typed claims tied to a DOI and either a verbatim quote or an explicit evidence locator. It currently covers 2.4 million claims from 147,000 papers. The paper says AskChem offers a web interface plus REST, SDK, and MCP access for AI agents. In AskChem-Bench, a GPT-5.5 reader grounded in AskChem reached 100% resolvable DOIs, versus 88.3% without retrieval. ArXiv · AI/CL/LG's note
The system converts papers into atomic, typed claims tied to a DOI and either a verbatim quote or an explicit evidence locator. It currently covers 2.4 million claims from 147,000 papers. The paper says AskChem offers a web interface plus REST, SDK, and MCP access for AI agents. In AskChem-Bench, a GPT-5.5 reader grounded in AskChem reached 100% resolvable DOIs, versus 88.3% without retrieval. ArXiv · AI/CL/LG's note
score 4