NousLab.org

The challenge

Scientific research is slow, fragmented and expensive. Researchers spend weeks reviewing literature, classifying studies and writing reports that could be generated in hours. Existing tools were either too generic or too expensive for small and medium research groups.

The client wanted a platform that would change the way researchers work: allowing them to find relevant papers in seconds, extract data automatically, detect contradictions between studies and collaborate in real time, all without sacrificing rigour or data privacy.

Our contribution

We joined the project from day one as technology partners, not suppliers. We proposed the multi-tenant architecture to allow multiple institutions to share the infrastructure with complete data isolation. We recommended a separate Python microservice for AI processes, which proved key to scalability.

One of our most valuable contributions was the 3-layer queue system: we designed a solution that separates tasks by weight and priority, allowing the platform to support up to 36 simultaneous licences without service degradation.

We also proposed and built the ethics committee module , a feature not in the original scope that has turned out to be a very strong competitive differentiator in the university sector.

What we built

• Multi-database semantic search: simultaneous connection to 12 scientific databases (PubMed, Semantic Scholar, CrossRef, arXiv and more), understanding the real meaning of a query, not just keywords.
• 3 autonomous AI agents: a paper analysis agent, a contradiction detection agent and a continuous monitoring agent for new relevant results.
• Intelligent data extraction: the system reads PDFs and automatically extracts figures, tables, statistical results, study populations and molecules in structured format.
• Ethics committee with 32 normative frameworks: AI assistant that guides researchers through the ethical implications of their studies, with configurable profiles per institution.
• Integrated CRM and billing: complete client management, subscription plans, automatic PDF invoice generation and payment tracking.
• Multi-tenant architecture: full data isolation per organisation, with granular licence, role and permission management.
• 3-layer queue system: robust management of heavy processes with Redis, ensuring AI tasks never block the user experience.

Technology

Results

NousLab.org is an actively live platform aimed at scientific research teams: universities, CROs, biomedical research teams, hospitals and pharmaceutical companies. Any organisation managing research with rigour and volume can get immediate value from the platform.

Technically, the project is highly complex: it requires orchestrating multiple AI agents in parallel, maintaining coherence across heterogeneous scientific databases, processing PDFs with structured data extraction, implementing multi-tenant logic with full per-organisation isolation and managing heavy-process queues that never block the user experience. All of this behind an interface that must remain accessible to researchers with no technical background.