Compare solutions

Choosing the right data anonymization tool

Compare and filter these anonymization solutions by feature.
Designed for DPOs, CISOs, DBAs and QA teams.

Select your criteria ↓

9 solutions compared

= not documented

DOT Anonymizer

  • Personal data discovery : documentedOver 100 detection rules, AI run on premises
  • Functional consistency : documentedFunctional consistency preserved across multiple sources
  • Unstructured documents : documentedFree text, PDFs with layout preserved, metadata and OCR
  • Deployment: On premise : documentedCloud : documentedOn premises or in the cloud
  • Data subsetting : documentedWith DOT Extract
  • API or CI/CD integration : documentedAPI and CI/CD integration via DOT Extract
  • AI integration : documentedAnonymization scripts generated from natural language; LLM of your choice (OpenAI, Mistral…), cloud or on premises
  • Multi-DBMS : documentedOver 30 technologies: relational DBMS, NoSQL, CSV, JSON and XML files
  • Dynamic masking : not documented
  • Synthetic data : not documented
  • Trial version : documentedFree sandbox trial
  • See 5 more criteria

Delphix

  • Personal data discovery : documentedSensitive data profiling included
  • Functional consistency : documentedFictitious data with referential integrity across systems
  • Unstructured documents : not documentedWhole file masking documented; no document formats such as PDF or Word supported natively
  • Deployment: On premise : documentedCloud : documentedVMware, OpenShift ; AWS, Azure, Google Cloud, IBM Cloud
  • Data subsetting : not documented
  • API or CI/CD integration : documentedREST API to drive masking
  • AI integration : documentedAI models run locally; AI-based synthetic generation (Delphix Synthetic Data, 2026)
  • Multi-DBMS : documentedDistributed databases, mainframe, PaaS and files
  • Dynamic masking : not documented
  • Synthetic data : documentedDelphix Synthetic Data
  • Trial version : not documentedDemo on request; no free trial documented
  • See 5 more criteria

Informatica Data Masking (IDMC)

  • Personal data discovery : documentedSensitive data classification in Cloud Data Governance and Catalog
  • Functional consistency : documentedRepeatable outputs and referential integrity across tables
  • Unstructured documents : not documented
  • Deployment: On premise : not documentedCloud : documentedIDMC SaaS platform: only the Secure Agent is installed in your IT environment
  • Data subsetting : documentedSubset extraction within the masking task
  • API or CI/CD integration : documentedREST API to run tasks
  • AI integration : documentedAutomatic classification of sensitive data by CLAIRE AI
  • Multi-DBMS : documentedHundreds of data sources via IDMC connectors
  • Dynamic masking : not documentedNot documented in IDMC; Dynamic Data Masking is a separate on-premises product, out of scope
  • Synthetic data : not documented
  • Trial version : documentedFree Cloud Data Integration trial
  • See 5 more criteria

Broadcom Test Data Manager

  • Personal data discovery : documentedPersonal data profiling and detection
  • Functional consistency : documentedConsistent masking across distributed, cloud and mainframe environments; cross-database consistency via HASHLOV
  • Unstructured documents : not documented
  • Deployment: On premise : documentedCloud : documentedWindows Server, Docker on Linux or Kubernetes on premises; Kubernetes deployment on Azure AKS, AWS EKS or Google GKE; no SaaS offering
  • Data subsetting : documented
  • API or CI/CD integration : documentedIntegration with DevOps pipelines
  • AI integration : documentedData Assistant Plus: AI-derived synthetic data generation rules
  • Multi-DBMS : documentedDB2, IMS, distributed databases
  • Dynamic masking : not documented
  • Synthetic data : documented
  • Trial version : not documented
  • See 5 more criteria

DATPROF

  • Personal data discovery : documentedAnalyze module (analyst source)
  • Functional consistency : documentedDeterministic masking across tables and databases; consistent unique values across databases since Privacy 4.23
  • Unstructured documents : not documented
  • Deployment: On premise : documentedCloud : not documentedDedicated server or virtual machine, Windows or Linux
  • Data subsetting : documentedSubset module (analyst source)
  • API or CI/CD integration : documentedMasking automation in CI/CD
  • AI integration : documentedOptional AI assistant, disabled by default, via the OpenAI (GPT) API and your own key: external engine
  • Multi-DBMS : documented
  • Dynamic masking : not documented
  • Synthetic data : documented
  • Trial version : documentedFree 14-day trial license
  • See 5 more criteria

IBM Optim Data Privacy

  • Personal data discovery : not documented
  • Functional consistency : documentedSame identifier masked identically across databases, for example Oracle and Db2
  • Unstructured documents : not documented
  • Deployment: On premise : documentedCloud : documentedIBM Optim 2.0 (July 2026): containerized application on premises, in a private cloud or public cloud
  • Data subsetting : documented
  • API or CI/CD integration : documentedPublished API reference
  • AI integration : not documented
  • Multi-DBMS : documentedDB2, Oracle, SQL Server, Informix, Sybase, Teradata…
  • Dynamic masking : documentedMasking at SQL query time via UDFs, and a masking API
  • Synthetic data : not documented
  • Trial version : not documented
  • See 5 more criteria

Tonic Structural

  • Personal data discovery : documentedSensitivity scan
  • Functional consistency : documentedConsistent subsets across linked databases
  • Unstructured documents : documentedVia Tonic Textual, a separate product
  • Deployment: On premise : documentedCloud : documentedStructural Cloud with a Professional license; cloud or self-hosted with an Enterprise license
  • Data subsetting : documentedProfessional or Enterprise licenses; API configuration in Enterprise only
  • API or CI/CD integration : documentedREST API that integrates with CI/CD pipelines
  • AI integration : documentedAI agent-assisted data subsetting
  • Multi-DBMS : documentedApplication databases, data warehouses, Spark, files; Db2 documented for LUW, not for Db2 for i
  • Dynamic masking : not documented
  • Synthetic data : documented
  • Trial version : documented
  • See 5 more criteria

Talend (Qlik)

  • Personal data discovery : documentedAutomatic detection of semantic types: email, phone, identifiers
  • Functional consistency : documentedRepeatable masking: same input and same seed produce the same value
  • Unstructured documents : not documented
  • Deployment: On premise : documentedCloud : documentedTalend Data Fabric, Studio and Remote Engine installed at the customer site; Qlik Talend Cloud as SaaS
  • Data subsetting : not documented
  • API or CI/CD integration : documentedCI Builder (Maven), CommandLine, Talend Cloud public API
  • AI integration : documentedAI functions in pipelines (Databricks, Snowflake), run on the target platform
  • Multi-DBMS : documentedHundreds of sources, including AS400 components
  • Dynamic masking : not documented
  • Synthetic data : not documentedSubstitution values generated in tDataMasking, without generating full synthetic datasets
  • Trial version : documented14-day Qlik Talend Cloud trial
  • See 5 more criteria

Xcase for i

  • Personal data discovery : documentedColumn groups identified by metadata, vocabulary and data
  • Functional consistency : documentedShared conversion dictionaries, including across heterogeneous databases
  • Unstructured documents : not documented
  • Deployment: On premise : documentedCloud : not documentedWindows client, collaborative mode on a central workstation
  • Data subsetting : documentedTest-DB: targeted subsets, parent-child integrity preserved
  • API or CI/CD integration : not documentedGenerated SQL script, repeatable and automatable; no REST API or CI/CD integration documented
  • AI integration : not documentedAI assistants documented for Xcase v12, the modeling tool, not for Xcase for i
  • Multi-DBMS : documentedNative Db2 for i; Anonymize-DB: DB2, SQL Server, MySQL…
  • Dynamic masking : not documentedRCAC is a native Db2 for i feature, independent of Xcase for i
  • Synthetic data : not documentedOnly mentioned in a 2018 Ithéis presentation; not covered in current documentation
  • Trial version : documentedOne-month trial or POC offered by Ithéis
  • See 5 more criteria

Short version: the 6 essential criteria. Get the full version to see all 11 criteria.

Get the full version

Compare all 9 solutions on 11 criteria: the 6 essentials, plus AI integration, multi-DBMS support, dynamic masking, synthetic data and trial version availability.

Instant access on this page once you fill in the form.

Full version

How to choose based on these criteria

There is no single best solution. It all depends on your databases, your regulatory constraints and how your teams populate their test environments.

Automatically identifying sensitive columns is the first step of any project. Without a reliable inventory, a single overlooked data item is enough to expose everything. Identifying your personal data

The same customer must receive the same substitute identity in every database; otherwise end-to-end tests fail on joins that no longer match.

Personal data does not live only in databases: PDFs, emails, reports, application logs. Check that the tool anonymizes these documents while preserving their layout, and consistently with the databases. Anonymizing unstructured text with AI

On premise, the solution is installed in your IT environment: production data that has not yet been anonymized never leaves it. In the cloud, it runs at a hosting provider or as SaaS: simpler to operate, but data hosting must be governed by contract.

Extracting a consistent subset of production reduces volumes, speeds up refreshes and limits the exposed surface. Discover DOT Extract

Triggering anonymization from a continuous integration pipeline lets you refresh test environments without manual intervention.

AI speeds up sensitive data detection, test dataset generation and rule configuration. Check where the models run: an external AI engine can send production data outside before it is anonymized.

An IBM i environment almost always coexists with Oracle, SQL Server, PostgreSQL or flat files. A single solution avoids maintaining duplicate rules.

Values are masked on display according to the user’s profile, without modifying the database. Useful in production, but it is not anonymization within the meaning of the GDPR.

Generating data from scratch complements anonymization for cases missing from production, but does not always reproduce real-world correlations. Synthetic or anonymized data: which method?

Testing on your own schemas remains the safest way to validate compatibility before a call for tenders.

Building your own anonymization tool? Feasible for a single database, riskier as soon as several sources must stay consistent. Read our comparison

The product cards were last updated on October 28, 2026, based on each vendor’s public documentation. This comparison was prepared by ARCAD from publicly available information, in particular the official documentation of the vendors concerned, as consulted on the date indicated. The solutions are compared against common functional criteria, selected for their relevance to the use cases presented. “Not documented” means that the public sources consulted do not confirm that the feature in question is available. It does not mean that the feature is missing. Features may vary depending on solution versions, modules, licenses and configurations. The information presented is subject to change, and the vendors concerned may report any error or change so that the comparison can be updated. Product names mentioned are the property of their respective owners.This page contains no affiliate links.

Spotted an error? Each card includes a “Report an error” link.