Welcome to HyperFlow’s powerful world of “Super Nodes” and service adapters! This guide will introduce you to one of HyperFlow’s most distinctive architectural features - the ability to use a single node type with multiple service implementations. This approach dramatically increases the power and flexibility of your AI applications while keeping the interface simple and consistent.
In HyperFlow, “Super Nodes” are flow graph nodes that leverage an abstract service architecture, allowing them to connect to multiple implementations of the same capability. Think of these nodes as versatile power tools with interchangeable attachments:
This means you can build a flow graph once and experiment with different AI providers or database technologies simply by changing a dropdown selection.
When you add a Super Node to your flow graph, you’ll notice a Service parameter in the node’s settings panel. This parameter allows you to select which specific implementation you want to use for that node’s function.
For example, the Call LLM node can connect to OpenAI, Anthropic, Google, or many other AI providers. Each service might have its own specific parameters (like model selection or temperature settings), but the node’s inputs and outputs remain consistent regardless of which service you choose.
This architecture offers several benefits:
Note that the list of services documented below are those available at release 1.0 and we expect an increasing number across all categories to be added over time, so please keep an eye out for future release notes and blog postings announcing new service adapters.
Further, we will be releasing shortly a plugin SDK allowing our users to develop both new service implementations for existing nodes as well as new nodes themselves.
HyperFlow organizes its Super Nodes into several categories. Here’s a guide to the main node types and the service implementations available for each.