← Back to blog Privacy

How we anonymize your data

"Anonymized" is one of those words that gets used loosely. Here's exactly what it means in our pipelines, and what we do to make sure no identifiable information ever leaves your systems.

What gets removed

Before any data is shared, every direct identifier is stripped at the source. That includes:

Indirect identifiers — combinations of demographic facts that could re-identify someone in a small dataset — are flagged and either generalized (age 60 instead of date of birth) or removed entirely.

Where the stripping happens

Anonymization happens inside your systems, before any data reaches us. We work with your existing pipelines and tooling rather than asking you to ship raw data to a third party for processing. If your institution already has a de-identification step in place, we slot in after it.

Nothing identifiable ever leaves your network. We never see the raw, identifiable form of any record.

What we actually receive

What ends up in our hands is structured records that look like clinical or experimental cases without any link back to a real person. For example, a patient record might become "65-year-old, elevated creatinine, CKD stage 3, prior hypertension." A lab experiment might become a list of conditions, instrument settings, and measured outcomes with no operator name or sample ID.

Compliance frameworks

The process we use is designed to meet HIPAA Safe Harbor standards in the US, GDPR's anonymization criteria in the EU, and the equivalent local frameworks where we operate. If your institution has additional requirements, we work them into the pipeline rather than around them.

Have specific privacy requirements you'd like us to meet? Tell us about them.