Two ways to run the engine.
Run the open engine on your own hardware, free. Or have NovoMCP set it up on your infrastructure, under a branded subdomain, with the compliance a regulated pipeline needs. Same engine underneath, on your infrastructure either way.
Self-hosted
Free forever
Run it on your own hardware, including an NVIDIA DGX Spark with Novo Station.
- The complete engine, MCP + REST server, 69 tools
- All ADMET / property prediction models
- All compute-service reference images (docking, MD, QM, NNP)
- Pluggable auth, credit metering, audit sink, LLM provider
- Docker Compose one-liner for a full local stack
- Weekly releases, GitHub issues + Discussions
Deployment service
Contact for pricing
NovoMCP sets it up on your infrastructure and keeps it current.
- Everything in Self-hosted, set up by NovoMCP on your infrastructure: your cloud, your cluster, or an NVIDIA DGX Spark
- A branded subdomain for your deployment (yourorg.novomcp.com)
- FAVES compliance API: regulatory status, structural alerts, BOILED-Egg, and prior art, with per-field provenance and an audit trail on every call
- FEP (free-energy perturbation) through the hosted API
- GPU services wired to your own GPUs (docking, MD, QM, NNP)
- Upgrades on the release train, transparent changelog
- A named NovoMCP engineer for setup, upgrades, and support
Not sure which fits? Self-hosted is right for research, teaching, prototyping, and any team that already owns its infrastructure. Deployment service is right for regulated pipelines that need compliance screening inline and a team to set the engine up. The engine is the same either way, and it runs on your infrastructure either way. The difference is who sets it up.
Install it yourself, or have NovoMCP set it up.
Request access to scope the fit for your team in a call.