A tool designed for legal professionals
A specialized neural translation engine for the legal sector, offering accuracy and reliability.
Layered architecture
- General training
The engine is first trained to translate generalist legal texts such as doctrine or articles, taking into account the terminological variations and nuances specific to each legal system.
- Specialized training
Once the basic training has been completed, the engine is duplicated, and each new engine is fed with specialized documents from a particular field of law. The general-purpose engine gives rise to several engines, each with its own specialty.
- Terminology training
Enrichment with knowledge of specific areas of law, such as commercial law, criminal law, civil law, etc.
Data origin
Lexa has been trained on millions of legal documents, covering a wide range of sectors and jurisdictions. This includes texts ranging from international contracts to national laws, court rulings and regulatory body regulations.
Legal documents
Over 6 million legal documents, compiled in 33 languages, have been rigorously selected for engine training: contracts, leases, GTCs, writs, decisions, statutes, RGPD clauses.
Legislation & regulations
Integration of civil codes, case law, tax regulations, international employment standards as well as legal acts including those of the EU (General Data Protection Regulation, Competition Directives), SEC (Financial Disclosure Rules) and other global authorities.
Regulatory texts
Lexa integrates documents from regulatory bodies such as the EU (General Data Protection Regulation, Competition Directives), the SEC (Financial Disclosure Rules) and other global authorities.
Human in the loop
Specialized lawyers
Lawyers from various fields contributing to the development of the engine.
Legal translators
Translation experts refine results for maximum accuracy.
Language engineers
Professionals optimize engine performance.
Data & ethics
To guarantee maximum accuracy, all the data collected in partnership with organizations has been the subject of considerable work.
For Lexa to integrate these official texts, the majority of them had to be checked by our linguists, corrected, cleaned of any unnecessary numbers, symbols or phrases and then anonymized.
After training the motor, several tests were carried out to find the best setting for Lexa.
Legal specialization
Lexa is entirely dedicated to the legal sector.
Optimum precision
Layered architecture and specialized data for reliable translations.
Adaptability
Customization to meet the specific needs of each user.
