Enterprise AI Platform
10x acceleration
1-minute
100% cost saving
What is an Enterprise AI Platform?
How the Platform Works?
Deployment and Pricing Options
Self-Hosted
Deploy our platform in your on-premises, private cloud, edge, or VPC.
Maintain full ownership of infrastructure.
Ensure control over your data and AI models.
Ideal for enterprises with GPU availability and those who are looking to manage the whole platform internally.
Managed Service
We install and maintain our platform in our partner's cloud for you.
Includes cloud GPUs.
Available in leading data centers in Iceland, Sweden, Finland, and Azure.
Ideal for enterprises that need an AI platform quickly and don't want to worry about hardware, installations, or updates.
All-in-One AI Stack for the Enterprise
Manager
Admin backend for the developer portal and orchestrating deployments.
- Handles creating pipeline-runs, authorization policies, user access roles, etc.
- All writes to the master DB (both user and application writes) are handled by this service.
Realtime
Listen/notify implementation for the master DB.
The developer portal and deployed services subscribe to updates for the services they have access to.
Admin portal
Admin user interface for managing the stack.
Non-admin users see an "AppStore" like grid-view of the end-user applications they have access to.
CI/CD
- Tekton core stack in the "tekton-pipelines" namespace.
- Allows building and deploying services from source code or image registry. It is used for all of our internal services and images, and for deploying services from the manager.
- If you make custom changes to pipelines or tasks, you have to run
helm upgrade stack-basebefore changes take effect. - A separate helm chart is used to define the infrastructure needed for services deployed by the CI/CD.
Traffic management and authz
Istio provides a service mesh for internal and external traffic management, security, and observability.
- For external communication, it's used for JWT token checks, mapping traffic to correct services, and handling TLS certs together with cert-manager.
- For internal communication, Istio provides mutual TLS and service discovery inside the stack.
- "Istio-system" namespace for Istio base resources deployment.
- "istio-ingress" namespace for gateway deployment.
- All namespaces with Istio-sidecar enabled will have Istio Envoy sidecar automatically deployed to enforce mTLS.
User management & authentication
- Keycloak provides a full user management and authentication service.
- It uses CNPG Postgre cluster as its database.
- Initialized with realm-export.json to create correct realms and clients on install.
Databases
- Stack uses CNPG operator that manages the deployment of Postgre clusters.
- Both the master database and Keycloak database are deployed as part of the stack-base helm chart using the operator.
- Capability to deploy separate Postgre clusters for applications with pg-vector and other extensions.
Cert Management
Allows automatic cert management using LetsEncrypt.
- For this, the stack needs a public IP and correct DNS setup.
Providing certificates manually is also possible.
Why Enterprises and Top Innovators Choose Us?
The platform lets you securely deploy production-grade AI systems (LLMs, chatbots, agents) that connect securely to your private data – boosting speed to market from months to days.
Unlike cloud-based AI services where you have to send your data to the cloud, your can run our platform in your infrastructure (on-prem, private cloud or VPC) so data never leaves your control.
All common sources – PDFs, databases (SQL/S3), file shares, plus custom sources using APIs
Anywhere Kubernetes runs, including on-prem servers, private or public cloud, edge, or hybrid environments. But we’ve abstracted Kubernetes complexities so you don’t even realize it’s the underlying infrastructure.
Anywhere Kubernetes runs – on-prem servers, private or public cloud, edge, or hybrid environments
AI has forced enterprises to reimagine their workflows, processes and how they work. The key to this new digital disruption is the speed of adoption. Because the worst-case scenario is a slow reaction which can result in a loss of competitive advantage.
Enterprise AI platforms increase the speed of adoption, because building technology infrastructure from the ground up is complicated and requires many specialized roles.
Our enterprise AI platform is designed with security in mind. We focus on the most demanding data security scenarios in on-premises and private cloud environments, so that developer building AI applications securely.
