See cross-domain risks and opportunities early — as evidence-labeled investment theses for review. Never advice, never automated trades.
Get candidate hypotheses and research leads with their evidence posture attached — and a citation when a claim is literature-supported.
Turn detected gaps into patent concepts and buildable opportunities, drafted for you to carry forward.
Spot operational stress and failure patterns across domains before they become public — with a human in every loop.
Run a domain from a review queue of surfaced problems and drafted work. You decide what gets pursued.
Signal is everywhere; meaning is not. Most systems score data inside a single domain. LIMEN Helix treats the patterns that break a market, a body, a supply chain, or an institution as variations of the same underlying dynamics — and maps them onto one shared structure.
It ingests 260+ public and institutional sources on a roughly 30-second refresh, scores stress and phase across 20 domains, connects them through a 123-region connectome, and surfaces what's worth a human's attention — with an evidence label on every claim.
Twenty domain models — each computing stress, phase, diagnoses, and opportunities from the same refreshed public-data snapshot, then passing cross-domain signals to one another. A shock in energy is felt in finance; a failure in supply chains surfaces in trade and infrastructure.
At the center is a connectome of 123 brain regions, networks, and molecular targets — a shared coordinate system. Each node binds to the domains it informs, so a pattern detected in one domain can be mapped, by structure, into another. This is how a discovery in neuroscience becomes a candidate hypothesis in medicine, or in markets.
This is the discipline that makes the scale trustworthy: every claim carries an evidence posture, and nothing is dressed up as more certain than it is. Nothing is discarded for being unproven, either.
Supported in the literature. Strengthen and cite. (Literature-supported — not clinical validation.)
A claim worth testing — a research target.
Not yet examined — a gap to explore. A queue, not a failure.
Contradicted — kept and studied for why it fails.
About 5% of mapped claims are literature-proven today, and the rest are labeled plainly. As verification works the queue, proven rises and not-yet-examined falls — in the open.
LIMEN surfaces candidate cross-domain treatment hypotheses — where an intervention that works in one field suggests an untried analog in another — and labels each by evidence state. Proven claims carry real citations linked to their source literature.
When the system finds something worth pursuing, it drafts the paperwork across five lanes. Drafts are drafts: buttons say "Draft Candidate Memo," not "File."
Nothing is filed, submitted, traded, applied for, or diagnosed automatically. The autonomous fire pipeline stays gated by design.
Every artifact is a draft for human review. The operator reviews, decides, and separately acts — every time.
Roughly every six hours the system checks its own integrity and applies only known-safe, mechanical fixes. Anything needing judgment is flagged for an operator — never silently changed.
LIMEN is built to be run by people. Twenty domains map to a future of domain operators — humans who review, decide, and pursue what the system surfaces. The engine finds and drafts; people judge and act. That is the design, not an afterthought.
Tell us how you'd like to engage. We read every message — no automated outreach, a person follows up.