Flows Overview

Build visual workflows that decide how your project responds to messages and supported connector events.

What are Flows?

Flows is a visual workflow builder inside each project. You connect nodes on a canvas to define what happens when a user sends a message or when a supported connector event arrives. Each project can have one active flow per connector (for example, one for GetCourse and one for Telegram).

Instead of one fixed reply path, you can combine knowledge-base answers, LLM prompts, conditions, variables, connector outputs, and supported GetCourse actions.

How a flow works

A flow is made from connected steps:

Start - a connector receives a message or supported event.
Process - connected nodes read data, call AI tools, set variables, or perform actions.
Choose a path - condition nodes can route the flow based on message content or saved values.
Reply or finish - output nodes send a response or complete the flow.

What you can build

AI answers

Search your project knowledge base (RAG) or invoke an LLM with custom prompts and variables.

Branching logic

Route by conditions, time windows, or message filters before choosing the next step.

Data transformation

Set variables, transform text, and reuse values from earlier steps.

Other services

Connect a flow to another service when you need an extra lookup or action.

GetCourse support

Read learner data, manage groups, add notes, move departments, and close support dialogs.

Review and debug

Use logs and debug nodes to understand what happened during a test run.

Connectors at a glance

Connector	Best for	Notes
Chatbot Widget	Website chat	Use when visitors chat through the FileBrain.Pro widget.
iframe	Embedded chat	Use when the chat is embedded into a page with iframe.
GetCourse	Support dialogs on GetCourse	Use learner messages and GetCourse dialog actions in a flow.
Prodamus.XL	Prodamus.XL events	Use supported Prodamus.XL connector events.
Telegram	Messenger bot	Use messages from your connected Telegram bot.

Node categories

The flow editor palette is organized into these groups:

Input - receive data from the connector trigger
Output - send the final reply back to the channel
Control - conditions, time checks, filters
Data - variables and text operations
AI - RAG search and LLM invocation
Integration - connect to other services or send Telegram messages from a flow
GetCourse - user data and dialog actions on GetCourse
Debug - tracing and notes for troubleshooting

Detailed node reference and configuration guides are coming soon.

Billing

Flow runs use platform credits. AI nodes such as RAG Search and Invoke LLM may add usage-based costs. Check your balance before enabling flows for real users.