Company
Medicine deserves better training data.
AI is about to sit beside every clinician. What it learns from decides whether it helps. Terrals exists to make that foundation complete, linked and trustworthy.
Not another dataset. The ground truth medicine's AI learns from — every layer of every patient, ready for every stage of training.
What we believe
The whole patient
A diagnosis is never one scan or one note. Models should learn from every layer, in the order it happened.
Privacy first, always
Identifiers stay inside the hospital. We would rather ship less than ship something unsafe.
Real outcomes
The best signal for clinical AI is what actually happened to the patient next.
Partners, not suppliers
Hospitals that work with us gain research, infrastructure and compliance — not just a cheque.
What we do
Eight ways we work with you.
From a single imaging cohort to a prospective, multi-omic study — or putting your own data and AI to work, privately and DPDP-ready.
Large-scale multi-modal
EMR, imaging, H&E and omics for the same patients, linked on one de-identified ID and aligned on a single clinical timeline.
Learn more →Prospective omics
Consented, freshly collected samples sequenced to your spec — NGS, WGS, bulk and single-cell RNA — with matched clinical data.
Learn more →Imaging
CT, MRI, X-ray, mammography, PET-CT and whole-slide pathology, paired with the reports written about them.
Learn more →RL environments
Clinical decision environments built from real longitudinal records, for training and evaluating medical agents.
Learn more →For hospitals
Research partnerships that turn routine care data into funded studies, infrastructure and publications — with privacy built in.
Learn more →DPDP compliance
Get your health data ready for India's Digital Personal Data Protection Act — assessment, consent, de-identification and DPO support.
Learn more →Data structuring
Scattered hospital systems mapped into one standard, coded model (OMOP, FHIR, SNOMED CT, LOINC), inside your own walls.
Learn more →Private AI deployment
Language and imaging models deployed on your own infrastructure, evaluated on your data. No patient data leaves.
Learn more →Privacy is the product.
De-identified at source
Identifiers are removed inside the hospital before data is released.
Consented collection
Prospective samples are collected under informed consent.
Ethics-approved
Studies run under institutional ethics committee approval, following ICMR guidelines.
DPDP-aligned
Processing designed around India's DPDP Act 2023.
Secure delivery
Encrypted, access-controlled, time-limited delivery.