What Is Flovaly Workflows, and When Should You Use It?

Flovaly Workflows is a visual canvas where you wire generation steps — a script, an image, a video, a talking actor — into a pipeline you build once and rerun with a single click, instead of repeating the same setup by hand in the Studio every time. Use the Studio for a one-off clip; switch to Workflows the moment you catch yourself running the same sequence of steps more than once, or need to fan one input out into several variations at once.

What the canvas actually is

Open Workflowsfrom the sidebar and start a new one: a dark, pannable grid canvas opens with an inline name field. Right-click or double-click anywhere to drop a node from a category picker, then drag from one node's output handle to another's input handle to connect them — handles are colour-coded by data type, and the canvas won't let you wire an incompatible pair together. Hit Run in the action bar and the whole graph executes as one background job: credits are deducted up front, the job goes into a queue, and each node updates live as it finishes rather than you waiting on one blocking request.

Which nodes actually generate something today

The node picker is organised into six categories — Generation, Inputs, Models, Presets, Tools, and Upload — and every node type the picker lists is real; what varies is which ones actually run when you hit Go. Image, Video, and Talking Actor generation nodes work fully, alongside Text Input and Upload as free input sources and Image Models / Video Models as override nodes for picking which underlying model a generation step uses.

LLM, Enhance Prompt, AI Video Editing, Tools (resize, concat, extract frame), and Presets are not live yet. They appear in the picker and can be dropped onto a canvas, but running a graph that includes one of them makes that specific node fail immediately with a “Coming soon” error — the rest of the graph still finishes around it. In practice that means there is currently no in-app way to have a node polish a rough prompt for you, or to stitch several finished clips into one continuous file; both are worth knowing before you plan a graph around them.

When Workflows earns its place over the Studio

The Studio wins for a single, one-shot generation — it's faster to open, configure, and run once. Workflows earns its keep once you're repeating a pipeline: the same script-to-image- to-video sequence for a new product each week, several Talking Actor scenes strung into one storyboarded ad (our multi-scene storyboard guide walks that build step by step), or one source branching into a few different renders to compare side by side. Workflows also surfaces every intermediate output, not just the final result — useful for debugging exactly where a chain went wrong instead of only seeing the last node's failure.

What a workflow run actually costs

Each node is billed at the same rate as its Studio equivalent, summed across the graph before the run starts. An Image node is 2 credits, a Video node is 1 credit per rendered second, and a Talking Actor node is a flat 8 credits regardless of script length. Text Input, Upload, and the two Model-override nodes cost nothing — only nodes that actually generate media get billed. A three-step script → image → video pipeline rendering a 5-second clip comes to 7 credits (2 for the image, 5 for the video), about $1.35 on the $29 Starter plan. The full per-generation breakdown that these node costs are drawn from is in our AI video ad cost guide, and how many credits you need per month covers monthly budgeting across a mix of Studio and Workflows runs.

Availability, limits, and what runs in the background

Workflows isn't restricted to a specific plan — any signed-in account with enough credits can build and run one; there's no Pro-only gate in the product today. There's no cap on how many nodes a single graph can hold or how many workflows an account can save, but runs themselves are rate-limited to 15 per 10 minutes per account. Once submitted, a run has 15 minutes to start processing and 30 minutes to finish before it's automatically failed and refunded — generous margins for anything short of an unusually large graph.

Frequently asked questions

What is Flovaly Workflows?

A visual canvas for chaining generation steps — image, video, talking actor, and more — into a reusable pipeline you build once and rerun, instead of repeating the same manual setup in the Studio each time.

Is Flovaly Workflows only available on Pro or higher plans?

No — it's available to any signed-in account with enough credits to cover the graph's node costs. There's no plan-level gate on the feature itself.

Does Workflows support batch-generating ads from a list of products?

Not today. Hitting Run executes the current graph once, producing one set of outputs — there's no way to feed in a spreadsheet or list of inputs and have it loop the pipeline across each row automatically. For now, “batch” means rebuilding or re-running the same saved graph with new inputs yourself, one run at a time, which is still faster than reconfiguring the Studio from scratch each time.

Do all the nodes in the picker actually work?

Most do — Image, Video, Talking Actor, Text Input, Upload, and the Model-override nodes are all fully functional. LLM, Enhance Prompt, AI Video Editing, Tools, and Presets currently appear in the picker but fail with a “Coming soon” error if you try to run them, so don't plan a graph that depends on any of those five yet.

Does a workflow run happen instantly, or does it queue?

It queues as a background job. Submitting a run deducts credits immediately and hands back a run ID; the graph then processes node by node, and you can watch progress update live rather than waiting on one long blocking request.

The fastest way to see whether your repeated setup is worth turning into a workflow is to build one — start with the Workflows canvas overview or jump straight into the $1 trial and wire up your first pipeline.

FAQ