FAQ
Frequently asked
questions
What is Humboldt?
Humboldt is a cross-domain research discovery platform. It indexes 1,500+ papers across biology, economics, physics, and computer science — not by topic or keyword, but by structural mechanism — and surfaces connections between fields that no domain-specific search would find.
How do we get access?
Request early access. We onboard research teams on a rolling basis and work through a real research challenge with you during the first session.
Do we need expertise in other disciplines?
No. You describe your problem in the language of your own field. Humboldt handles the translation. Every result comes with enough context to evaluate relevance without needing background in the source discipline.
How is Humboldt different from Google Scholar or semantic search?
Keyword and semantic search find papers that use similar language. Humboldt finds papers that describe the same mathematical structure — regardless of vocabulary, field, or era. A 1960s thermodynamics result can match a 2010s network economics paper because the underlying dynamics are equivalent.
Who uses Humboldt?
Research teams, doctoral students, and independent researchers who routinely encounter problems that feel solved — just not in their field. If you suspect that the hard part of your problem has already been worked out somewhere, Humboldt is designed to find it.
How does Humboldt find structural equivalences?
Each paper is embedded into a vector space that represents its underlying mechanisms — not its surface content. Queries are mapped into the same space. Results are ranked by vector similarity, which captures how closely two descriptions share a causal or mathematical structure, regardless of the words used.