About Lattice
Built by Early Readout
The problem
Every year, thousands of university labs produce technologies that could become companies, save lives, or reshape industries. But finding them is nearly impossible.
Technology transfer offices at 200+ US universities each maintain their own listing portals, with different formats, different search tools, and no cross-institutional index. A deep tech investor scouting for licenseable mRNA delivery platforms has to manually visit MIT, Stanford, Johns Hopkins, Columbia, and dozens more, one by one.
Preprints on bioRxiv and medRxiv signal what researchers are working on months before patents are filed, but nobody connects preprint activity to licenseable technologies. The result: VCs and corporate R&D teams discover university IP too late, or not at all.
How Lattice works
Lattice aggregates technology listings from 200+ university tech transfer offices and millions of bioRxiv/medRxiv preprints into a single intelligence platform. Every record is processed by a multi-stage AI pipeline that extracts structured metadata, generates search-optimized representations, and computes three-dimensional scores: Novelty, Validation, and Opportunity.
Search combines multiple retrieval strategies: semantic understanding captures what a technology means, while lexical matching catches exact identifiers like patent numbers and gene names. A proprietary fusion layer merges these signals to surface results that neither approach finds alone.
Deep Dive lets you go further. Click any technology to open an AI-powered research session. Ask questions, explore related preprints, check patent status, and understand the field landscape. Every answer is grounded in real data, not speculation.
300,000+
Records
3M+
Preprints
200+
Universities
<2s
Search time
What makes it different
- Cross-institutional search. No other tool lets you search 200+ university TTO listings in one query. The data aggregation itself is the moat.
- 3D scoring. Every technology gets a Novelty, Validation, and Opportunity score. Sort and filter by what matters to you. Not just search, but screening.
- Deep Dive. Click any technology and ask questions. Get AI-powered answers grounded in preprints, patents, and publication data.
- Preprint-to-patent connection. Lattice links millions of preprints to university technologies, surfacing research momentum before IP is filed.
Where this is going
Lattice is the foundation for a predictive translation engine. The long-term goal is to track translational velocity: the acceleration from academic research to commercial intent. When a lab publishes a preprint, files a patent, and lists a technology for licensing within 18 months, that is a signal. We want to detect that signal automatically, 12 months before anyone else notices.
Today: search and intelligence. Tomorrow: alerts, field-level dashboards, and commercialization predictions. The data infrastructure is built. Now we are making it useful.