Luvia
Enormous AI spend goes into building biomedical knowledge graphs — and almost all of it is private or paywalled, rebuilt from scratch behind every company's walls and stale within a year.
Core trick
Stores typed claims with source-verbatim quote checks at write time, and derives standing and contradictions on read rather than storing or editing them.
Description
Enormous AI spend goes into building biomedical knowledge graphs — and almost all of it is private or paywalled, rebuilt from scratch behind every company's walls and stale within a year. Meanwhile researchers everywhere are already reading the same literature with LLMs. OpenClaims starts from that waste: if your agent is analyzing a paper anyway, it might as well log what it found, in a structure a crowd can actually query. Every reading makes the next person's question better answered — the opposite of a graph that dies in one lab.
The user we built for is any researcher whose agent already reads papers. That agent files what it found — typed (subject, predicate, object) claims with polarity, provenance, and the exact grounding sentence — through an MCP server, and the commons adjudicates. Nobody edits the graph; the graph re-adjudicates
itself.
Three guarantees make it trustworthy: quoted spans are checked verbatim against the source, so fabrications are rejected at write time by non-LLM code; standing is recomputed on read and never stored, so a wrong claim is outweighed in the open rather than silently deleted; and no model decides what's true — entity resolution is deterministic and contradiction is structural.
We seeded it with 494 verified readings across 25 papers on the microglial cGAS–STING–interferon axis. Out of that corpus it flags 11 genuine contradictions (e.g. STING→NLRP3 asserted positive in one paper, negative in another) and composes 25 cross-paper chains stated in no single source — for instance, Huntingtin → (↑cGAS, PMID:32581130) → (↑type-I-interferon, PMID:37095396) → an emergent HTT→interferon link neither paper makes, grounded verbatim in both. These are candidate hypotheses the structure surfaces, not validated findings — and the commons labels them exactly that. That honesty is the point: it tells you where the literature already disagrees and where a testable connection is hiding, across papers that never cite each other.
README excerpt
# OpenClaims **An open, living claim commons for the scientific literature.** OpenClaims is an append-only store of *claims* — typed `(subject, predicate, object)` assertions carrying polarity, provenance, and certainty. A researcher's agent reads a paper for its own purposes and files what it found through an MCP server; the commons adjudicates. **Nobody edits the graph. The graph re-adjudicates itself.** The primary record is a log of **readings** — one contributor's agent writing up what it found in one source, each claim carrying the exact sentence that grounds it. The claim table, corroboration, contradiction, standing, reading-fidelity, and every chain are *views* over that log, recomputed on read and disposable. Nothing in the log is ever merged, mutated, or deleted. Three properties fall out of that, and they are what make the commons worth contributing to: - **A quoted span is checked, verbatim, against its source.** A span that is not in the registered text is rejected as a fabrication — by non-LLM code, at write time. - **Standing is recomputed on read, never stored.** Retraction is a consequence of evidence, not a capability: a wrong claim is outweighed, demoted in the open, and stays on the record with its reasoning inspectable. - **No model decides what is true.** Entity resolution is deterministic, contradiction is structural, and standing is a lexicographic tuple you can predict and audit. What accumulation then buys you on the read side: | | |