For AI agents: a documentation index is available at the root level at /llms.txt and /llms-full.txt. Append /llms.txt to any URL for a page-level index, or .md for the markdown version of any page.
Logo
AI Hub
OverviewApp editorFlow editorAdminAPI & SDK
OverviewApp editorFlow editorAdminAPI & SDK
  • Flow editor
    • About flows
    • Creating flows
      • Working with flows
      • Planning flows
      • Integrating flows
      • Creating advanced apps from flows
      • Logging and troubleshooting
    • Flow step reference
    • Using custom functions
    • Flow guides
    • Reviewing flow results
AI Hub
On this page
  • Two ways to add upstream and downstream connections
Flow editorCreating flows

Integrating flows

Was this page helpful?
Built with
Enterprise Single-tenant

After you’ve created and compiled a flow, integrate it into your existing business processes and systems.

Flow integration occurs in two stages:

  • Ingesting documents into a flow, referred to as upstream integration.

  • Sending results at the end of a flow, referred to as downstream integration.

For example, in a mortgage processing flow, users upload documents to a mortgage lending application, which then sends the documents to your app for processing (upstream integration). Then, processed results are sent on to a loan processing application (downstream integration).

Two ways to add upstream and downstream connections

You can configure upstream and downstream connections in one of two ways:

  • Within the flow editor — Add logic directly in the flow using events. Configure pre-flow (upstream integration) and post-flow (downstream integration) custom functions.

  • Through an app deployment — First publish the flow as an advanced app. Then create an app deployment and configure your upstream and downstream integrations through the UI.

Use the flow editor approach when you want integration logic embedded directly in a flow definition. Use the deployment approach when you want operational integration settings managed outside the flow itself. The latter approach is useful when you want to reuse the flow logic with multiple deployments.