Custom AI App Control API Reference

HyperFlow AI also provides ready-made interactive chatbots that you can embed in your host applications or web-sites, and these chatbots also use this Control API. See the Embed code for these in the published flow-graph manager.

There is also a Chatbot Simulator, available via the Simulator mode button in the HyperFlow IDE that can be attached to a ready-to-publish flow-graph for testing that shows a detailed log of the API calls between flow-graph and AI app, which can be helpful in understanding the Control API in use.

Finally, there is a video walkthrough of the Control API on our Youtube channel.

Introduction

The Control API provides comprehensive access to a published flow-graphs functionality, allowing it to drive and interact with custom chatbots, AI assistants and other AI applications you might choose to develop to provide custom user-interfaces to your HyperFlow apps.

The control API can launch multiple, concurrent sessions of a flow-graph, each maintaining its own run context. This is typically used to enable the implementation of services such as end-user facing chatbots on a website, each user interacting with their own flow-graph session.

This documentation covers the available API calls and their returned properties. If you have any questions, please visit the HyperFlow Discord support server or email [email protected].

Getting Started

Required HTTP Headers

All Control API requests must include these headers:

{
  "Content-Type": "application/json",           // Required
  "Accept-Language": "en-US,en;q=0.9",          // Required for i18n middleware
  "X-API-Key": "hf_..."                         // Required for external requests
}

⚠️ Critical: The Accept-Language header is REQUIRED, even if you don't need internationalization. The HyperFlow i18n middleware processes this header on every request, and missing it can cause internal server errors. Always include a default value like "en".

Example Request:

fetch('<https://hyperflow-ai.com/api/flowgraph/control/start>', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Accept-Language': 'en',           // Don't forget this!
    'X-API-Key': 'hf_VUujE...'
  },
  body: JSON.stringify({
    flowGraphID: "...",
    queryParams: {}
  })
});

Additional Recommended Headers:

Publishing your production-ready flow-graph

Once you are satisfied with your flow-graph's development and wish to use it through either one of HyperFlow's built-in embeddable chatbots, or a chatbot or other custom UI you have created, you will need to publish it using the Publish button in the main HyperFlow editor UI. Only published flow-graphs can be control and interacted with via the Control API documented here.