ESM3 design
Generate and iterate masked-sequence or design-loop candidates, then force review before wet-lab handoff.
Protelynx.ai
Working ESM3, ESMC, and ESMFold2 workflows in 2-4 weeks, with auditable run records, responsible-use gates, reports, and reusable product extraction.
For research teams
Protelynx is a paid implementation lane for teams that have protein-model questions but need operational software: provider routing, batch runs, audit-ready records, review gates, exports, and repeatable reports.
The public proof stays honest. It uses mock output without credentials, then switches to Biohub or local/Hugging Face execution only when the customer provider lane is configured.
Model lanes
Generate and iterate masked-sequence or design-loop candidates, then force review before wet-lab handoff.
Triage FASTA sets, cluster candidates, and produce high-throughput embedding reports for teams with many sequences.
Create structure-confidence reports and exported artifacts when the question needs fold-level review.
Keep hosted Biohub and local/Hugging Face execution behind explicit credentials while the public page uses mock proof.
Launch Sprint
Choose one customer workflow, one input format, and one report that would matter within 30 days.
Wire CLI/API execution, provider routing, persisted run records, safety findings, and artifacts.
Ship a mock-provider report first, then run Biohub or local provider paths only after credentials are configured.
Export the customer bundle, extract reusable deltas, and decide whether the account converts into Workflow Hub.
Conversion handoff
payment link ready. Payment link ready.
Recurring review queue, templates, exports, reports, and implementation support after the sprint is accepted.
Analytics scripts are disabled until GA4 or PostHog public keys are configured.
Proof library