Sarus
Paris privacy layer that lets analysts query sensitive data and also generates synthetic data
Quick Overview
| Company | Sarus Technologies SAS |
|---|---|
| Category | Synthetic Data Platforms |
| Headquarters | Paris, France |
| Founded | Not stated |
| EU Presence | Yes - France (EU) |
| Data Location | Your cloud or on-premise |
| Open Source | GitHub organisation sarus-tech (licences not verified) |
| GDPR Compliant | Yes |
| Self-Hosting | Yes, Docker or Kubernetes |
| Main Features | Differential privacy, query rewriting, output-level control, synthetic data, SarusLLM, clean room integration |
| Pricing | See website; no prices published |
| Best For | Teams that want analysts to work on sensitive data without seeing it |
| Replaces | Gretel, NVIDIA NeMo Data Designer |
Detailed Review
Sarus presents itself as a privacy layer rather than a pure generator. Data scientists connect to a dataset they cannot read directly, and results come back with privacy protection applied.
Its legal notice names Sarus Technologies, a société par actions simplifiée registered in Paris.
Privacy before generation
Query rewriting turns a processing job into a privacy-safe version on the fly, and output-level control restricts what leaves. Synthetic data is one of several routes, produced by a generative model trained with differential privacy.
Where it runs
The site states that Sarus can be deployed on-premise or in a public cloud with Docker or Kubernetes, and lists integrations with PostgreSQL, Snowflake, MongoDB, Tableau, SageMaker and Vertex AI. A recent notice covers Azure Confidential Clean Rooms.
LLM workflows
SarusLLM is described as a way to use and fine-tune open-source language models with private data, which stretches the product beyond tables.
Limitations to Consider
Synthetic data is only one feature, so teams that just need a table generator may prefer Syntho or Dedomena. No pricing is shown and licensing of the GitHub components was not verified. Query-based access works differently from handing out a synthetic file, which changes how analysts work.
Who Should Use Sarus
Analytics, security and compliance teams that need controlled access to sensitive data with differential privacy guarantees rather than only a synthetic copy.
Pros and Cons
Pros
- Differential privacy at its core
- Deploys on-premise or cloud
- Query rewriting and output control
- Wide integration list
- GitHub presence
Cons
- Synthetic data is a secondary feature
- No pricing shown
- Open-source scope not stated
- Different workflow from file-based generators
Alternatives to Sarus
Other European synthetic data platforms worth comparing before you decide:
Octopize
France · Avatar anonymisation with synthetic data and public price tiers
Syntho
Netherlands · Masking, rule-based and AI-generated data in one engine
Dedomena
Spain · Madrid platform pairing validation, anonymisation and synthetic data
Aindo
Italy · Trieste synthetic data for clinical evidence and public health
See all European synthetic data platforms or compare them directly against Gretel on our alternatives page.
Frequently Asked Questions
Yes, from a generative model trained with differential privacy, though it is one of several routes on the site.
On-premise or in a public cloud, using Docker or Kubernetes, per the home page.
The site links a GitHub organisation but does not say which components are open source.