Emmy (CNRS) vs Charlie: Which AI Tool for Research?

Emmy: The CNRS General-Purpose Chatbot

Emmy is the result of a partnership between CNRS and Mistral AI, signed in October 2025. The main argument for this choice is digital sovereignty: facing an AI market dominated by American (OpenAI, Google) and Chinese players, CNRS opted for a European solution. User data is hosted in European data centers, compliant with GDPR and the EU AI Act.

Emmy's features:

• Text translation in all languages
• Document summarization and rephrasing assistance
• Web search
• Text and image recognition
• "Reasoning" mode for step-by-step question processing
• Adding external documents to contextualize responses

Emmy already has over 7,000 active users. Notably, Emmy's deployment comes with a ban on using other general-purpose AI tools (ChatGPT, Gemini, etc.) in a professional context at CNRS. Emmy does not use user prompts and documents to train its models.

Charlie: The Specialized AI Scientific Agent

Charlie takes a fundamentally different approach. Rather than a general-purpose chatbot, it is an autonomous AI agent designed specifically for biomedical research. The distinction is crucial: a chatbot answers your questions; an agent plans and executes complex tasks autonomously.

Charlie's features:

• Real-time access to PubMed (35M+ articles), PMC (10M+ full texts), GEO (6M+ genomic datasets), Espacenet (140M+ patents)
• Systematic citations with DOI, PMID, and patent numbers
• Multi-document analysis (hundreds of PDFs simultaneously)
• 7 scientific use cases: literature monitoring, article summaries, hypotheses, experimental protocols, results interpretation, publication preparation
• Deep specialization in biomedical research
• Personal workspace to organize your documents

Charlie is also 100% sovereign, hosted on certified French data centers, GDPR-compliant, and never reuses data for training.

Comparison Table: Emmy vs Charlie

Chatbot vs Agent: A Fundamental Difference

The most important difference between Emmy and Charlie lies not in their individual features, but in their fundamental architecture. Emmy is a chatbot: it answers questions one at a time, reactively. Charlie is an agent: it can break down a complex objective into subtasks, plan a strategy, query multiple sources simultaneously, and adjust its approach based on intermediate results.

Concrete example: You ask "What are the promising biomarkers for early diagnosis of Alzheimer's disease?"

• Emmy (chatbot): Generates a response based on its pre-trained knowledge, potentially outdated. Can perform a general web search if you explicitly ask.
• Charlie (agent): Automatically queries PubMed for recent publications, PMC for systematic reviews, GEO for gene expression data, identifies the most cited biomarkers, analyzes associated clinical studies, and synthesizes a comprehensive answer with 15-20 verifiable references (DOI/PMID).

"Emmy is an excellent tool for administrative tasks and general productivity at CNRS. But for actual scientific research, we needed a tool that understands our specific needs and can directly query scientific databases. That's why we use Charlie for our literature reviews." — Researcher, Institute of Biology, CNRS

Who Should Choose Each Solution?

These two solutions are not necessarily in competition: they address different needs. A CNRS researcher could use Emmy for administrative and communication tasks, while using Charlie for their actual scientific research. The key is choosing the right tool for each task.

Conclusion: Two Visions of Scientific AI

The emergence of Emmy at CNRS and Charlie illustrates a broader trend: French research is embracing AI while prioritizing sovereign solutions. These two approaches represent two complementary visions:

• Emmy democratizes access to AI for all CNRS staff across all disciplines, with a high-performing general-purpose tool.
• Charlie offers deep specialization for biomedical research, with an agent architecture enabling autonomous complex tasks.

The choice between the two depends on your specific needs. For intensive scientific use requiring verifiable sources and access to specialized databases, Charlie offers capabilities that a general-purpose chatbot cannot match. For daily productivity and varied tasks, Emmy fulfills its role perfectly.