Traversa · AI for biology, public health, and travel

Three products. One mission.

Each product solves a specific bottleneck in how biology, health, and travel data get turned into decisions.

Coming Q3 2026

Traversa Travel

Duty-of-care intelligence for study abroad.

AI destination intelligence for university study abroad offices. Replaces 40+ hours per semester of manual safety briefings with ISO 31030-compliant, audit-ready documentation in 15 minutes.

For: study abroad directors, risk officers

Status: in development, customer discovery

Notify me when live →

LIVE NOW

NativeReady

Predict native MS suitability from a protein sequence.

Paste a protein sequence. Get a calibrated suitability score and risk factors in seconds, before you waste days preparing samples that won't ionize. v0.4 model combines ESM-2 protein-language embeddings with 47 hand-engineered features (physicochemical + N-glycosylation sequon counts + transmembrane-helix prediction). Trained on 635 proteins from public literature, the RCSB PDB, and EuropePMC. Membrane-topology and glycosylation risk rows surface model-aware signals with targeted experimental recommendations. Out-of-distribution warnings when the model is uncertain.

For: native mass spec researchers, biotech R&D

Status: v0.4 with glycosylation + TM features, free during beta

Python SDK: pip install nativeready

Open benchmark: 635-protein dataset, code, and model on GitHub

Preprint: bioRxiv 2026.05.03.722506

Try free → Read the preprint →

RESEARCH PREVIEW

Traversa Pathogen

Automated viral surveillance reports from FASTA files.

Upload a viral genome (FASTA). Get an automated surveillance report with QC, lineage assignment when possible, genomic features, and publication-ready figures. Built around foundation models for virus prediction. Starting with arboviruses.

For: public health labs, academic researchers

Status: research preview, joining waitlist

Why Traversa exists.

Biology and public health generate enormous amounts of data. Most of it sits in databases nobody reads, papers nobody cites, or instruments that need a PhD to operate. We build tools that translate that data into decisions a real person can act on.

Honest by default

Every Traversa tool tells you what it doesn't know. Out-of-distribution warnings, calibrated confidence intervals, transparent limitations. We'd rather say "we're not sure" than give you a wrong answer confidently.

Built on public data

UniProt, PDB, GISAID, WHO, CDC, ProMED. The reference datasets of public health and biology are open. We use them, cite them, and contribute back.

For working scientists

Our users are bench biologists, study abroad directors, and public health analysts. Tools should be paste-in, paste-out. Quick to try, easy to interpret, and useful in the same hour you find them.

About.

Traversa is being built by a researcher working on AI for public health and infectious disease.

Brhanu F. Znabu

Co-Founder & CTO · Traversa

PhD Student · Biomedical Engineering · University of Nebraska-Lincoln

My research is building foundation models for virus prediction. Traversa grew out of a simple observation: the same signals that power my research can become audit-ready intelligence for the institutions that send people, samples, and resources across borders.

Turn complex bio-data into clear decisions.