AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis

AskChem is a chemistry literature search and synthesis system that indexes atomic claims rather than whole papers.

It uses a faceted taxonomy, evidence graph, and living taxonomy to surface typed claims with linked provenance.

The system indexes 2.4 million claims from 147,000 papers and is exposed through web, REST, SDK, and MCP access.

In benchmarks, grounding a GPT-5.5 reader in AskChem improved DOI resolvability and citation density.

Its main value is helping people and AI agents find evidence, verify sources, and synthesize findings across papers more reliably.