Legal · data processing agreement
DPA template — partner institutions.
Last updated 2026-05-12 · Effective 2026-05-12 · Counsel-reviewed full text linked at signature time. This page is the plain-English shape of the agreement.
Who this applies to
Vet schools, veterinary referral hospitals, and pharma BD partners sending us canine clinical or genomic data, or accessing the SciRouter API as an institutional customer.
What we do with partner data
Partner-submitted canine clinical and sequencing data is processed for the specific case workflows the partner authorizes — neoantigen ranking, vaccine construct generation, comparative-oncology atlas queries, and the partner's own aggregate reporting. We never use partner data to market to your patients.
De-identification standard
Owner identifiers and direct patient identifiers are removed at intake. Hash of the case ID + partner-side mapping table stays with you; we never see the mapping. Industry-standard re-identification testing applies on aggregate data products.
Co-publication terms
Standard collaboration agreement governs authorship + data-sharing rights for joint publications. Default: partner contributes clinical narrative + interpretive authorship; we contribute methods + reproducibility authorship.
Sub-processors
Current sub-processors: AWS (storage), Railway (orchestration), RunPod (GPU inference), Postgres provider, Sentry (error tracking), Resend (transactional email). We maintain an up-to-date list and notify partners at least 30 days before changes.
Term + termination
Initial term: as negotiated in the partner agreement, typically 12 months. Termination for material breach: 30 days to cure. On termination, partner-shared data returned or destroyed per partner preference. Aggregate-derived models continue to function but no new partner data is ingested.
Contact
Negotiate or request the full text: use the contact form and mention "Partnerships" in the message.