Grenoble-based plagiarism and AI-content detection platform used by 1,100+ institutions in 50+ countries; LMS-integrated and GDPR-compliant, with 4M+ students reached — though AI-detection reliability remains a documented limitation.
Founded in Grenoble in 2004, Compilatio SAS provides plagiarism-detection services to more than 1,100 educational institutions across 50+ countries, with particularly deep penetration in French and Italian universities. Over one million users and more than four million students have been made aware of academic integrity requirements through the platform.
Activities
The Magister+ tier, released in 2023, adds AI-generated content detection alongside traditional similarity analysis, comparing submitted documents against a corpus of billions of web pages, scientific publications, and institutional documents. The service integrates natively into Moodle, Blackboard, and other major LMS platforms, reducing administrative friction for faculty.
Results
Evidence of direct learning impact is limited: the platform's primary value is protective (deterring and detecting plagiarism), not instructional. Leading French institutions including the University of Lorraine declined to adopt the AI-detection feature in 2025 following published guidance warning of unreliable false-positive rates — a limitation Compilatio itself acknowledges by framing its AI-detection score as probabilistic, not proof. For non-native speakers and ESL students, false AI-detection positives carry documented equity risks. The core plagiarism-detection product, now 20+ years mature, remains well-evidenced through adoption and GDPR-compliant.
Implementation
Indicative cost
Low (< €50k)
Time to results
Short (< 1 year)
Staffing & skills
Compilatio SAS (vendor, product and support), Institutional IT/LMS administrators for integration
Conditions for success
Native integration into major LMS platforms (Moodle, Blackboard) to reduce faculty friction
Framing AI-detection scores as probabilistic guidance rather than definitive proof
GDPR-compliant data handling
Common failure modes
Unreliable false-positive rates led the University of Lorraine to decline the AI-detection module in 2025
Documented equity risk of false AI-detection positives for non-native/ESL students
Founded in Prishtina in 2016, Crossplag built one of the first detectors targeting cross-language ('translate-to-evade') plagiarism, reaching 5,000+ educators before …