Systran

What Is Systran?

In 1968, when the internet did not yet exist and the most powerful computers filled entire rooms, a company called Systran was founded in Paris with a single mission: to enable communication across language barriers using machines. That founding predates Google by thirty years, DeepL by nearly fifty, and the transformer architecture that underpins modern neural MT by half a century. Systran did not just participate in the history of machine translation; it wrote most of it.

The company spent its first decades providing translation technology to the US Air Force, the European Commission, and NATO. It built the first commercial machine translation systems used by governments to process classified intelligence across language barriers. When the European Union needed to communicate across nine, then twelve, then twenty-seven official languages, Systran's technology was a core part of the solution. This institutional DNA, built on deep expertise in structured, high-stakes document translation for organisations where errors carry real consequences, still defines what Systran does best.

In January 2024, Systran was acquired by ChapsVision, a French sovereign data processing group whose clients include major public-sector and defence organisations. Operating as "SYSTRAN by ChapsVision," the company now sits within a broader ecosystem of French data intelligence tools oriented around data sovereignty, a significant strategic alignment given that Systran's strongest selling points have always been on-premise deployment and strict data privacy. For European enterprises and governments evaluating machine translation in 2026, Systran is no longer merely a historical curiosity; it is a viable, actively developed platform with uniquely strong privacy credentials.

Key Features

On-Premise Deployment for Full Data Sovereignty

The feature that most clearly distinguishes Systran from DeepL and Google Translate is on-premise Server deployment. With Systran Server, the entire neural translation engine runs within the customer's own infrastructure. Documents submitted for translation never leave the organisation's network. There is no cloud round-trip, no API call to an external service, no retention of data on Systran's servers. For defence contractors, intelligence agencies, law firms handling privileged communications, and regulated financial institutions, this matters enormously. It is the difference between a viable solution and one that the legal or compliance team will reject.

No other major commercial machine translation provider at Systran's quality level offers on-premise deployment as a standard product. This is a genuine competitive moat, not a marginal feature.

Custom Neural Model Training

Beyond raw translation quality, the most important variable for professional translation is domain accuracy: whether the MT system correctly handles the specific terminology, abbreviations, and style conventions of a particular field. Systran's Premium and Enterprise plans allow organisations to train custom neural models on their own translation memories, glossaries, and terminology databases. A pharmaceutical company can train Systran on clinical trial documents; a law firm on contract language; an engineering firm on technical manuals.

Custom model training produces measurably better results for specialised content than any general-purpose model, including DeepL's domain-agnostic approach. For high-volume document translation where consistency and accuracy matter, this capability justifies the higher Enterprise pricing.

GDPR-Compliant Cloud with Non-Retention Policy

For organisations that can use cloud-based translation, Systran's cloud infrastructure is hosted in Europe and enforces a strict non-retention policy: translated content is not stored, not logged, and not used to improve Systran's models. This is explicitly stated in Systran's data processing terms and distinguishes it from some competitors where the data handling policy requires careful scrutiny.

The ChapsVision acquisition reinforces this position. ChapsVision's entire business proposition is built around data sovereignty for French and European public-sector clients, and a data-mishandling incident at Systran would be catastrophic for ChapsVision's core market. The incentive alignment is clear.

CAT Tool and Enterprise System Integrations

Systran integrates natively with the major Computer-Assisted Translation (CAT) tools that professional translators and language service providers use: SDL Trados Studio, memoQ, and Across Language Server. It also integrates with Microsoft Office applications (Word, Excel, PowerPoint), SharePoint, ServiceNow, and email systems. The REST API supports integration with virtually any enterprise application.

For language service providers and large in-house translation teams, this means Systran can function as the MT engine within existing translation management infrastructure, providing machine-assisted throughput without disrupting established workflows.

Pricing

Systran Translate Pro offers cloud-based plans billed monthly or annually (annual billing includes a discount of up to 30%):

• Free — Limited text translation via the web interface; no document translation; no API access
• Lite — €6.99/month; unlimited text translation; document translation (Word, Excel, PowerPoint, PDF); user dictionary; 1 user
• Plus — €19.99/month; everything in Lite plus API access, advanced glossary management, priority processing; up to 3 users
• Premium — €44.99/month; everything in Plus plus custom neural model training, translation memory, team collaboration features, and priority support

Enterprise / Server — Custom pricing for on-premise Systran Server deployment, which includes dedicated hardware configuration, custom model training at scale, SLA guarantees, and dedicated support for government and high-compliance environments.

A Developer API tier with pay-per-character pricing is also available for technical integration use cases.

EU Compliance and Privacy

Systran's compliance posture is one of the strongest in the machine translation market. The cloud service is hosted in Europe and processes translated content under GDPR. Systran's non-retention policy (translated content is not stored and not used for model training) means the data protection risk profile is lower than general-purpose AI tools that log queries.

For the most sensitive use cases, on-premise Server deployment eliminates cloud data risk entirely. HIPAA-compatible configurations are available for healthcare use cases where translated medical content must remain within regulated infrastructure.

Since January 2024, Systran operates within ChapsVision's sovereign data group, which specialises in building technology compliant with French national security classification requirements. This institutional context makes Systran one of the few commercial MT providers that can credibly bid on classified government translation contracts in France and, by extension, other EU member states with similar sovereignty requirements.

Data Processing Agreements (DPAs) are available for enterprise customers who require documented GDPR compliance accountability.

Who It's Best For

Government agencies and defence contractors requiring machine translation with full data sovereignty will find Systran's on-premise Server option uniquely suited to their needs. No other commercial MT provider at this quality level offers production-grade on-premise deployment.

Enterprise translation teams and language service providers handling high volumes of specialised documents (legal, medical, technical, financial) will benefit from custom neural model training, which produces significantly higher accuracy than general-purpose models on domain-specific terminology.

Compliance-conscious European enterprises that need to translate documents internally and cannot send content to US-based cloud services will find Systran's EU-hosted cloud with non-retention policy a defensible choice that a data protection officer can approve.

Less suited for individual consumers or small teams who want a polished, fast consumer experience. DeepL's interface is markedly cleaner, and its translation quality on everyday text is excellent. Also less suited for global enterprises needing 130+ language pairs; Google Translate's breadth exceeds Systran's 55+ languages for truly multilingual operations.

The Verdict

Systran is a product with a unique position in the machine translation market that its history earns and its current ownership reinforces. If your primary requirements are consumer-grade translation with a polished interface and excellent quality on everyday text, DeepL is the better choice. But if you need on-premise deployment, custom domain model training, or a provider whose EU data sovereignty credentials are beyond question, Systran sits in a category largely by itself among commercial MT platforms. The ChapsVision acquisition has given it stronger institutional backing and a clearer market positioning than it has had in years. For European enterprises and governments with real compliance requirements, it deserves serious evaluation alongside the better-known alternatives.

Frequently Asked Questions

Is Systran GDPR compliant?

Yes. Systran's cloud service is hosted in Europe and complies fully with EU GDPR requirements. A key data protection principle Systran enforces is that translated content is neither stored nor used to train or improve its models, so your data remains your property. Systran provides Data Processing Agreements (DPAs) for enterprise customers and supports right-of-erasure and data portability requirements. For organisations with the strictest data sovereignty needs (defence contractors, government agencies, healthcare providers), Systran also offers fully on-premise Server deployments where translations never leave the customer's own infrastructure.

How does Systran compare to DeepL?

Systran and DeepL take different approaches. DeepL offers a cleaner, more modern consumer interface and superior translation quality on everyday text, particularly for Western European languages. Systran's strengths lie in enterprise and institutional use cases: custom neural model training on proprietary terminology, on-premise deployment, and deep integrations with CAT tools like SDL Trados and memoQ. Systran also has far more history in specialised domains such as defence, legal, and technical translation. For a small professional team needing polished translations, DeepL is often the better choice. For enterprises requiring data sovereignty or custom domain models, Systran offers capabilities DeepL does not.

Can Systran be deployed on-premise?

Yes. Systran Server is an on-premise machine translation solution designed for large organisations and governments that cannot send documents to external cloud services. With Systran Server, the entire translation engine runs within the customer's own infrastructure, so translations never traverse the internet or touch Systran's servers. This is particularly important for classified government data, regulated financial information, and proprietary legal documents. On-premise deployment also allows custom model training with confidential terminology without any data leaving the organisation. Server licensing requires direct engagement with Systran's enterprise sales team.

What languages does Systran support?

Systran supports translation across 55+ languages with more than 150 direct language pair combinations. Supported languages include all major European languages (English, French, German, Spanish, Italian, Dutch, Portuguese, Polish, Russian, Ukrainian, and more), plus Arabic, Chinese (Simplified and Traditional), Japanese, Korean, Hindi, and many others. Systran has particular depth in European language pairs due to its long history serving EU institutions and European governments. Custom neural models can be trained for any supported language pair with domain-specific terminology, further improving accuracy for specialist content.

Who owns Systran now?

Since January 2024, Systran has been owned by ChapsVision, a French technology group specialising in sovereign data processing and analytics. The acquisition was announced as part of ChapsVision's strategy to build a comprehensive suite of data intelligence tools for enterprise and government clients. Systran continues to operate under its own brand as 'SYSTRAN by ChapsVision' with Vincent Godard leading the Systran division. ChapsVision is headquartered in France and serves primarily public-sector and defence clients, giving Systran stronger access to institutional contracts that require French sovereignty guarantees.