neomundi.org

What It Is

NeoMundi Research is a French loi 1901 non-profit association positioned as an independent public measurement infrastructure for generative systems. It is not a chatbot, writing tool, or model platform. Instead, it attempts to observe generative AI services as “social objects”: without depending on the LLM providers being observed or selling products, it publishes public runtime mappings, anonymized data, observable logs, technical publications, and reproducible datasets.

Core Capabilities and Methodology

Its core components include the Law E™ theoretical framework and related measurement instruments. The page clearly states that its measurements observe only three types of objects: generated outputs, temporal dynamics, and public execution environments. It does not access model internals, does not reuse observed content, does not use processed data for training, and semantic processing is transient, serving only runtime measurement. This makes it closer to AI governance, external auditing, and research infrastructure than to a traditional AI tool.

Pricing and Commercial Boundaries

NeoMundi Research states that it “does not sell anything.” Its research is published as open access, data is released under CC-BY 4.0, and source code uses the MIT license. If operators want to measure their own AI systems, commercial operations are handled by the separate entity Louis M Sàrl. The page does not disclose plans, pricing, payment methods, or trial policies, so the actual procurement threshold remains unclear.

Pros and Cons

Its strengths lie in a clear statement of independence: it has no financial relationship with the LLM publishers being evaluated and separates public research from commercial operations. Its methodology, data, and technical materials emphasize openness and reproducibility, and its privacy principles are also relatively clear. The limitations are that it does not judge whether answers are true or false, but only publishes stability signals; its mapping results are also described as phased and partial snapshots. In addition, the page lacks product entry points, APIs, dashboards, customer cases, and support information, making it difficult for ordinary users to get started directly.

Who It Is For and Access from China

It is better suited to AI governance researchers, public institutions, compliance audit teams, model operators, and organizations focused on generative AI observability. It is not suitable for individual users looking for ready-made AI productivity tools. Access from China is not discussed in the text, and network connectivity, payment methods, and localization support are all unknown. If alternatives or complementary tools are needed, evaluation and observability solutions such as HELM, Chatbot Arena, OpenAI Evals, LangSmith, TruLens, and Ragas are worth following.