BLEACH

De-identification as an API, not an install.

Call Bleach from the pipeline you already run. Send regulated records in, get synthetic replacements back. No platform to adopt, no software to install.

Bleach mascot, a bleached-white capybara illustration

De-identification API, Call it from your pipeline, Configurable per pipeline, Smart replacement, Independent privacy assessment

The problem

01

Sensitive data supports research, AI, analytics, and data sharing, but identifying information limits where teams can use it.

Bleach brings de-identification into the pipelines teams already run. Call the API, choose what to detect, and control how the output is handled.

See it work

Input: raw recordA1
From: Jane Doe <jane@acme.com>Re:   Tuesday dock delayCall me at 432-555-0187 if thebay is still closed.
Output: bleachedA2
From: Priya Okonkwo <priya64@example.net>Re:   Tuesday dock delayCall me at 512-440-2291 if thebay is still closed.

Configurable per pipeline. Tune which entities you catch and how smart replacement rewrites them.

Illustrative output

Where teams use it

Three use cases

  1. 01

    Licensing training data

    De-identify a dataset for external licensing on your terms. Configure the API to match what your buyer requires, not a fixed template.

  2. 02

    Internal model fine-tuning

    Give ML teams working data without exposing identifying information. Call the API from the training jobs you already run.

  3. 03

    Data partnerships

    Move faster on cross-company data deals by de-identifying records through the API before they leave your pipeline.

How it fits

An API you call from the pipeline you already run. Nothing to install, nothing to operate.

Request the technical overview →
01

Call it from your pipeline

Send records from the jobs you already run. Bleach returns synthetic replacements in the same flow.

02

Configurable per pipeline

Choose which entity types to detect, then have Bleach replace, redact, or tag the matches. Smart replacement can preserve naming style and organization type.

03

Independent privacy assessment

Available and signed by an independent reviewer where required.

Access

EARLY ACCESS,
BY DESIGN.

Sign-ups open before general availability. Here's what happens after you request access.

  1. 01We'll reach out to set up a short call.
  2. 02We'll scope your pipeline and the data you need to de-identify.
  3. 03You get API access once we've confirmed it's a fit.

Request early access

Integrations

Already live on Databricks Marketplace. Run Integral's privacy risk assessment directly inside your existing workspace.