Licensing
Acuratio Enterprise
The federated learning platform. Multiple organizations train machine learning models collaboratively without centralizing data — per-organization annual license, three deployment modes.
Acuratio Enterprise lets multiple organizations train machine learning models together on data that physically stays where it is. Each participant runs its own node; the model learns across all of them without any raw record ever crossing an organizational boundary.
That data-stays-put property is what makes regulatory compliance manageable: when records never leave their custodian, the legal and contractual surface around sharing collapses, and the platform can be operated in regulated, audited, or even air-gapped environments without changing the underlying methodology.
Licensing model
Per organization
Each organization that joins the platform keeps control of its own data and nodes.
Annual subscription
12-month terms, automatically renewable. Product updates included throughout validity.
Privacy-focused
Data is never shared between organizations, making regulatory compliance easier.
Deployment options
Pick the modality that fits your sovereignty, infrastructure, and operations requirements.
| On-Premise | PaaS | Clean Room (SaaS) | |
|---|---|---|---|
| Data location | Customer infrastructure (bare metal, VM, private cloud / VPC); air-gapped supported | Customer cloud account (AWS, GCP, Azure) | Acuratio multitenant |
| Managed by | Customer | Customer team | Acuratio |
| Infrastructure | Docker, Podman, Kubernetes; air-gapped images; local files / S3 / GCS | Managed Kubernetes (EKS, GKE) or AWS Fargate; S3 / GCS / Azure Storage | Acuratio-hosted, elastic scaling; storage managed by Acuratio |
| Onboarding | Variable | Variable | Under 24 hours |
| Best for | Strict data residency, regulated workloads | Cloud-native teams that want sovereignty without on-prem ops | POCs, evaluations, fast starts |
What's included
Every Acuratio Enterprise license gives access to the full platform capability set.
Federated learning
- FedAvg with anti-poisoning reputation system
- FedGBT (federated XGBoost), horizontal and vertical
- Split Learning, FedGLM, Multi-Task Learning, federated K-means
- SHAP interpretability and federated cohort generation
Privacy by design
- Private Set Intersection (PSI) and negativization
- k-anonymity and variants (α,k / l / c,l-diversity)
- Integrated differential privacy
- Diffie-Hellman secure aggregation and Shamir Secret Sharing
Remote analytics & governance
- Split Dataset Pipeline
- Drift monitoring
- Granular access control and full traceability
Connectivity & networking
- VPN mode
- AcuProxy mode
- Connectors: S3, GCS, local files (CSV, Parquet)
- REST APIs, gRPC, Server-Sent Events
Support & updates
- Technical support per contracted SLA
- Product updates during the subscription
- Security patches included
- Technical documentation and deployment guides
Pricing adjusts to deployment mode, number of nodes, and any associated professional services. Tell us your use case and we’ll come back with a tailored proposal.