Pricing

An endpoint is a service in the stack that you can connect to using an API key and a URL. It can be a simple model inference or a more complex RAG (Retrieval-Augmented Generation) or agent system.

Each endpoint has a unique ID that is part of the URL path (for routing requests to the correct service). Most of these services use OpenAI-like chat completion APIs. However, depending on the type of endpoint, for example, in RAG systems, there is also a way to upload files to the system, and those files will be included in the chat completions.

We have partnered with leading data center providers around the world to host our platform in the following locations: Iceland, Sweden, Finland (coming soon), and Microsoft Azure.

Our managed service, available with the Scale license, is where we help you install, manage, and maintain our platform in our partners' cloud environments. So, you don’t have to worry about purchasing hardware, fixing issues, or performing updates.

With the Enterprise license, this option is also available at an additional cost. In that case, we will help you set up and manage the platform in any environment.

Our platform helps you build and deploy complete AI systems not just LLMs.
‍
Unlike others, we do not just provide you with a platform and leave you to figure out how to build your first RAG system, semantic search or other applications. We provide everything you need to create these assets. This includes one-click deployment of LLM endpoints, databases, and storage. So you can launch your first PoC in minutes, not months.

Also, with secure data connectors, you can connect these systems to your private data securely, and with simple APIs bring their capabilities to your offerings or legacy tools.

The platform supports most common file formats such as PDFs, HTML, plain text, and repositories. It also integrates with SQL databases, Windows (SMB) files and S3 storage through API connectivity. Additionally, you can add custom data sources using API code.

You can deploy it anywhere you can run Kubernetes: on-prem, on your existing virtual machines, bare-metal servers, public cloud, private cloud, or VPC.

Note: Developers don't need extensive Kubernetes experience, as we have removed its complexities. You won't even realize it's the underlying infrastructure.

Book a demo, and our team will show you how the platform works and how it can help with your use case.