Neuronautix · Services & Products
Neuronautix works in three complementary modes: scientific consulting for teams who need expert help on HCM, behavioral analysis, and FAIR metadata; R&D co-development for partners who want to build new methods, algorithms, and tools; and productized offerings — built once, deployed many times — that package our research into something a team can buy, install, or subscribe to.
R&D co-development
Through our flagship collaboration with the LIRMM lab (CNRS, Montpellier) led by Dr. Konstantin Todorov, Neuronautix runs active research workstreams that generate novel tools and publishable methods. Companies and consortia can join these as co-funders, co-developers, or licensees.
Using the Home-Cage Monitoring Ontology (HCMO) and related vocabularies to constrain LLM reasoning over preclinical knowledge graphs — increasing efficacy, reducing hallucination, and enabling provenance-aware scientific Q&A.
Developing embedding methods that combine neuroscience priors and statistical structure to generate new scientific knowledge from preclinical behavioral and physiological datasets.
LangGraph-based scientific agent squads with human-in-the-loop checkpoints, applied to literature synthesis, evidence mapping, metadata curation, and Virtual Control Group construction.
Service areas
We combine behavioral neuroscience expertise, Home-Cage Monitoring knowledge, FAIR metadata strategy, and practical software development to help laboratories move from fragmented experimental outputs to interpretable, reusable research assets.
Independent support for comparing, selecting, deploying, and documenting commercial, DIY, video-based, sensor-based, and multimodal Home-Cage Monitoring platforms.
Workflows for behavioral tracking, physiological recordings, ECG and HRV analysis, circadian metrics, longitudinal activity, and multimodal alignment.
Practical metadata schemas, provenance structures, dataset documentation, and FAIR-ready workflows for Findable, Accessible, Interoperable, and Reusable preclinical data.
Custom scripts, notebooks, data transformation tools, project prototypes, dashboards, and open-source workflows designed around real laboratory constraints.
Ontology-aware metadata structuring, semantic mapping, controlled vocabularies, SHACL-style validation concepts, and knowledge graph preparation.
Training and live webinars on Home-Cage Monitoring, behavioral analysis, metadata management, FAIR principles, data-care culture, agentic tools, and AI-assisted scientific workflows. See training topics & webinars.
Training & webinars
Neuronautix delivers interactive training sessions and online webinars — for individual labs, core facilities, consortia, and conferences. Each can be run as a focused webinar, a hands-on workshop, or a tailored in-house program. Topics span FAIR metadata, data-care culture, and practical AI for everyday research.
Designing and applying practical metadata schemas, provenance structures, and ontology-aware documentation so preclinical and behavioral data become Findable, Accessible, Interoperable, and Reusable in real laboratory workflows.
How rigorous data stewardship directly serves the 3Rs: better metadata, reuse of historical controls, and reproducible analysis reduce wasted animals and make every study count — making data welfare an ethical, not just technical, concern.
Building team habits, roles, and shared norms around documenting, validating, and curating research data — moving from one-off compliance to a sustained culture where caring for data is part of doing good science.
A practical introduction to AI agents and multi-agent workflows for science: what they are, where they help (literature synthesis, evidence mapping, metadata curation), and how to keep humans in the loop with provenance and validation.
Hands-on training on using Claude Code to accelerate day-to-day research tasks — data wrangling, analysis scripts, reproducible pipelines, and documentation — safely and effectively within a working lab.
Workflow
Each project is scoped around the scientific question, available systems, data quality, analysis needs, metadata requirements, and the level of transfer expected by the research team.
Deliverables
Neuronautix focuses on deliverables that make research work easier after the project ends: clear comparisons, reproducible analyses, structured metadata, readable documentation, and practical training material.
Tools and systems
Projects may involve Home-Cage Monitoring platforms, behavioral tracking tools, physiological recordings, metadata tables, ontology resources, custom Python workflows, and open-source scientific software prototypes.
Products
Our open-source research and ontology work feeds three productized offerings. Each one packages a piece of our R&D into something a team can buy, install, or subscribe to without a custom engagement.
Product · Local-RAG knowledge stack
A turnkey local deployment of a multi-agent LLM stack with retrieval over your internal corpus — SOPs, study reports, regulatory documents, code, slide decks, lab notebooks, partner contracts. Runs on-prem or in a private cloud. No data leaves your perimeter.
Built on the same multi-agent architecture as Neuronautix's Science Agent Squad, with an ontology-aware retrieval layer (HCMO, MBO, NAMO) that makes domain queries far more precise than naïve vector search.
Product · FAIR Virtual Control Group reports
Send us a historical control dataset (CSV, study reports, telemetry exports). We return a FAIR-compliant Virtual Control Group report: matched historical controls, statistical comparability checks, provenance trail, and the metadata package needed to defend the choice in a regulatory submission.
Built on the open-source FAIR-VCG pipeline, with LLM-assisted curation and ontology-anchored metadata to keep every claim traceable to a source record. Aligned with 3Rs animal-replacement principles and emerging NAM / IND data expectations.
Product · Ontology support contracts
The Home-Cage Monitoring Ontology (HCMO) and Mouse Behavior Ontology (MBO) are open-source semantic vocabularies maintained by Neuronautix and collaborators. For technology vendors, platform owners, and consortia, we offer paid support contracts on top of the open base.
Why these three
Neuronautix runs active R&D in three areas — multi-agent scientific reasoning, ontology-constrained LLM navigation (with LIRMM, Dr. Konstantin Todorov), and neuro-AI / stat-AI embedding methods for preclinical knowledge generation. The products above package the most mature, deployable outputs of that research. Buying a product also funds the next iteration of the open-source stack.
Start a project
From Home-Cage Monitoring to FAIR metadata, reproducible analysis pipelines, and deployable products, Neuronautix helps you transform complex experimental data into reusable scientific knowledge.
