AI image workflows for shops, marketers and creators

Build an image workflow once. Run it on every product.

Chain AI models and editing steps on one canvas: cut out, restage, resize for every channel, add your headline. Then run it on a spreadsheet of 500 products, or publish it as a simple app for your team.

  • Local models free every day
  • gpt-image-2 included
  • Your own AI keys, unmetered
A green ceramic mug on a grey backdropproduct.jpg
The mug cut out on a transparent backgroundcut out · 12 s · free
The mug on a light oak table in morning sunoak table, morning · gpt-image-2
One flow, three steps. Change the photo and run it again.

How it works

  1. 1

    Lay out the steps

    Drag steps onto the canvas and connect them, start from a template, or describe what you want ("cut out my product, put it on marble, make story and square versions") and FlowGen lays it out for you.

  2. 2

    Run it and tweak

    Run one step or the whole flow. Every step shows its pictures, what it cost and how long it took. Change a prompt and only the steps after it run again.

  3. 3

    Run it on everything

    Paste a spreadsheet and run every row as a batch, download a ZIP, publish the flow as a one-button app for your team, or call it from your code.

Start from a template

See all templates

The right model for each step

On our GPU, free

FLUX.1 schnell and DreamShaper 8 for pictures and edits, RealESRGAN for upscaling, InSPyReNet for clean cut-outs. Licensed for commercial use.

gpt-image-2, included

OpenAI's newest image model, best at text in pictures and precise edits. Three a day free, more on paid plans.

Every model on OpenRouter

Nano Banana, GPT-5 Image and whatever comes next, listed live with the price per picture before you run. Paid from cloud credit.

Your own keys

Add an OpenAI, Google or OpenRouter key and use those models without FlowGen limits. Keys are stored encrypted and never shown again.

Free

$0

  • 40 local GPU steps a day
  • 3 gpt-image-2 pictures a day
  • 25 saved flows, 1 published app
  • Batches of 5 rows
Create free account

Pro

$12 / month

  • 400 local GPU steps a day
  • 10 gpt-image-2 pictures a day
  • Unlimited flows, 10 apps
  • Batches of 100 rows
  • $3 cloud credit every month
Get Pro

Studio

$29 / month

  • 2000 local GPU steps a day, first in the queue
  • 25 gpt-image-2 pictures a day
  • Unlimited apps, batches of 500
  • The API
  • $10 cloud credit every month
Get Studio

Prices include tax. Cloud credit packs from $5 for paid OpenRouter models. Your own AI keys are never metered.

Questions

What is FlowGen?

A canvas where you chain image steps (create, edit, cut out, upscale, resize, add text and logos) into a workflow you can run again on any picture, on every row of a spreadsheet, or as a simple app for other people.

Do I need to know AI or design?

No. Start from a template or describe what you want in a sentence. Every step has plain settings, and "Improve" rewrites a rough prompt into a good one.

Can I run one flow on hundreds of products?

Yes. Paste a spreadsheet into the flow's data table, write prompts like "studio photo of a {color} {product}", and run the batch. You get one result per row and a ZIP of all of them.

Which models can I use?

FLUX.1 schnell and DreamShaper on our GPU for free, gpt-image-2 within your daily allowance, every image model on OpenRouter with cloud credit, and your own OpenAI, Google or OpenRouter key without limits.

What is a published app?

A page with a small form built from your flow: you pick which fields people fill in (a photo, a product name), they press one button and get the result. Their runs use their own allowance, not yours.

Who owns the pictures?

You do. They stay private in your gallery until you share a link, and you can delete them, or everything FlowGen holds about you, at any time.

Make the flow once. Let it do the rest.

Start free