🤖 Agent Mode

Agent Mode for Dataflow is a powerful solution for building a render-on-demand production pipeline.
Batch rendering is already powerful, but Agent Mode fully automates your rendering process.
It behaves like a watch folder, but for data.
It periodically scans your connected data source for new input and automatically starts rendering when available.
It can run 24/7 on a dedicated rendering machine, providing an ideal render-on-demand solution for your Adobe After Effects projects.

This opens a world of possibilities!
A common use case is an Asset Order Form (e.g., Google Forms) filled out by a customer with specific preferences for headlines, colors, brand elements, format, length, etc.
This data lands in the connected data source, and the Agent automatically generates the requested video based on your templated After Effects project.

Info

The Agent scans for new data every 15 seconds.

Getting the Most Out of Agent Mode

A really big benefit of Agent Mode is that it can monitor more than one data source at a time!
For example, in your Google Sheets selection, you can add as many Google Sheets as you want to your connected sheets selection.
This allows you to make the most of your rendering infrastructure.
Say you have multiple campaigns that are all connected to a front-end user fill-out form.
Each of these campaigns has very different needs of input data, render settings, output destinations, etc., so they all need a different Google Sheet suited to the needed parameters.
Without the ability to connect multiple data sources to Agent Mode, you would need to have multiple machines and Dataflow licenses set up to monitor each data source and render when new data comes in.
That's why connecting the Agent to more than one data source is so powerful, allowing you to squeeze everything out of your rendering Agent🤖.

On the other hand, you can also run multiple Agent instances on the same data source to distribute heavy rendering workloads.

Note

To manage rendering priority, the Agent scans connected data sources from top to bottom, processing the first source with new data before moving to the next.