Small models.
Big precision.
Our AI Core is a system of specialized models, all designed to perform specific tasks. We don’t ask one giant brain to do everything. We have visual experts, semantic experts, legal experts, comparison experts and generation experts. They work together to augment your analysis so you go 10x faster.
It’s about data.
You can’t learn IP law from Wikipedia or ChatGPT. We train our models on our proprietary European Lawcase Database. We don’t just cover the basics. We ingest data from the EUIPO and EPO for pan-European rights, the WIPO for international context, and every major National Office (INPI, DPMA, OEPM, etc.).
Our models have absorbed over 200,000 decisions—from Boards of Appeal rulings to CJEU jurisprudence and national court verdicts. We then fine-tune this raw legal intelligence with experts to ensure absolute predictability.
2 patents. 2 theses. 4 scientific publications.
Hosted in OVHCloud datacenters.
2 years of dedicated R&D
GDPR & AI Act ready.
Fully European.
Documented.
Sovereign.
Proven.
Our logic for trademarks.
AI Core is not a single model. It’s a coordinated system where each model specializes, contributes and reinforces the overall understanding of trademarks.
Item & Semantic Matcher
It compares goods, services, and claims semantically. It detects real similarity, identity, inclusion, and overlaps, going far beyond simple keywords.
Juridic Translator
It translates content with legal precision. It ensures complex IP nuances are perfectly understood across languages, based on real legal decisions.
Visual, Conceptual & Phonetic
It creates a multidimensional analysis. It sees the sign, understands the concept, and analyzes how the mark sounds (aurally).
Caselaw Retriever
It fetches the proof. It identifies and pulls the specific case laws relevant to your conflict to ground the analysis in reality.
Argument Generator
It doesn’t just score; it drafts. It turns data and jurisprudence into structured, defensible arguments ready for your review.
Speed is a feature.
AI Core combines data processing, multi-modal embeddings, reasoning channels and evaluation layers into one coherent flow. The result: fast inference, interpretable results and consistent behavior across thousands of trademark cases.
Explore what’s beyond AI Core.
Lipstip’s AI Core is only one part of a larger, interconnected ecosystem.
Behind every output, every similarity score, every interpretation, every insight. There are complementary technologies that reinforce accuracy, transparency and speed.
To understand how our system works end-to-end, we invite you to explore the modules that extend, challenge and enrich the AI Core: the algorithms that compare, the datasets that train, and the architectures that bring everything together.
Each component has its own role. Together, they form the technological backbone that makes Lipstip reliable, sovereign and ready for the next generation of IP analysis.
Triple Matching Algorithm
Get the best results with our own Triple Matching Algorithm.
Lawcase Database
A complete and proprietary database for your use.
Our tech
Combining all our systems, we are building an unique IP tech.