Akhi's Achilles
Drug-resistant infections kill ~1.27 million people a year, and the hard part isn't naming a resistance gene — it's finding the reversible opening resistance leaves behind.
Core trick
Deterministic code computes statistics while the LLM only retrieves and cites, and claims lacking provenance are dropped from a hash-chained evidence graph.
Description
Drug-resistant infections kill ~1.27 million people a year, and the hard part isn't naming a resistance gene — it's finding the reversible opening resistance leaves behind. I built Achilles, an evidence-grounded discovery console. Point it at your data and it builds a provenance-checked graph where every claim carries a citation: a deterministic Python core does all the math, and Claude only reads literature, retrieves, and cites — it never invents a number. If a claim can't be grounded, it doesn't become an edge.
The result is science you can break on stage. Against 29 independent public controls, Achilles recovers known biology (12/12), refuses an adversarial battery of plausible falsehoods (17/17), and fabricates nothing (0) — writing the whole verdict to a tamper-evident, hash-chained ledger anyone can re-verify. On a time-split hold-out it even anticipates relationships before the confirming paper appeared.
Shown end-to-end on antimicrobial resistance, it names the reversible target resistance creates, folds it with AlphaFold, and proposes a cited antibiotic-cycling hypothesis — bench to bedside, never medical advice. Domain-agnostic, reproducible from public data, and callable as MCP tools from any Claude agent. Honestly, watch my demo, but more so just use the product on your own, and trust me, you will see the technical biological depth this product has to offer.
README excerpt
<p align="center"><img src="frontend/public/logo.svg" alt="Achilles" width="340"></p> <p align="center"><b>An evidence-grounded discovery console — grounded answers you can re-verify, not take on trust.</b></p> <p align="center"><em>Deterministic core · provenance on every edge · reproducible from public data · MIT</em></p> <p align="center"> <a href="https://achilles-science.vercel.app">Live app</a> · <a href="https://achilles-science.vercel.app/methods">Methods</a> · <a href="https://achilles-science.vercel.app/bridge">Research ⇄ clinic bridge</a> </p> --- Point Achilles at your data — strains, variants, a literature set — and it builds a **provenance-checked evidence graph**: every claim carries a citation, a deterministic core does the math, and the language model only reads, retrieves, and cites. Nothing is asserted without a source, and the whole validation result reduces to a **tamper-evident hash-chained ledger you can re-verify**. The console is **domain-agnostic**. Antimicrobial resistance (*Burkholderia multivorans*) is the worked example, loaded behind a "Demo data" toggle — not the product's identity. ## The one idea **Claude proposes, code decides.** A deterministic Python core computes everything quantitative — lineage reconstruction, flipper (allele-reversal) detection, the target rank score, the collateral-sensitivity math. The model does exactly two things: extracts typed claims from literature, and narrates already-computed results with citations. I