The engine behind real discovery.
NovoMCP is the computational chemistry engine for drug discovery and materials science. One engine. Two domains. Four surfaces: chat, desktop, browser, and editor.
The bottleneck is the work between the science, not the science.
Drug discovery takes twelve years because of the work between the science: every handoff between tools, every compliance check deferred to the end, every week waiting for compute.
The engine pre-computes the knowable, automates the routine, and keeps the researcher at the center of every decision that matters. It checks in when judgment is needed. The rest runs.
Four principles, held to.
The engine is the product
Models are interchangeable. The engine they run on stays. NovoMCP is the layer that stays when the model is swapped: the 69-tool catalog, the ADMET and clinical-outcome models, the FAVES compliance layer, the immutable audit trail, and cross-run memory.
Open by default
The engine, wrappers, surfaces, and all trained models ship under Apache-2.0. The FAVES compliance API is a hosted service, access on request. What is open and what is hosted is stated up front, so a first-time reader knows what they get and what needs to be scoped.
Keep the researcher at the center
The 11-stage funnel runs autonomously and checks in when judgment is needed. Human-in-the-loop is where the science happens.
State the limits
The engine is signed and notarized, local-first, with a SOC 2 audit still ahead. Formal validation is in progress. Prediction is weaker on oncology and CNS. The boundary of what the engine can do is published. That boundary is what makes the rest credible.
NovoMCP is operated by Quant NexusAI Inc. Products: Novo (cloud platform) and Novo Compute (GPU and quantum compute).
Build on the engine, not around it.
NovoMCP is open to a small group of PIs, postdocs, and research engineers. Tell us what you are working on.