Turn an Idea into a Design

Post Ideas isn't a list you copy from. Every idea is one click away from a finished design. Here's how the handoff works.

Create a visual from an idea

1. Open Post Ideas and click an idea card to open its detail panel.
2. If the idea needs something only you have (a quote, a stat, a photo), you'll see a Needs your input badge and the inputs listed at the top of the panel. Fill them in and click Save inputs. The Create button stays disabled until they're all filled, since a testimonial post without the testimonial isn't much use.
3. Pick a model in the model picker on the Post Ideas page if you want something other than your brand's default.
4. Click Create Visual.

Krumzi bundles the idea's title, description, visual direction, and your text inputs into one brief and sends it to the AI Designer. You're taken straight into the editor and the AI starts building. The idea is marked Already created so you can tell what's done.

What format you get

Each idea carries a format, and Create Visual respects it:

• Image ideas open as a square 1080×1080 design.
• Carousel ideas open as a square carousel with the idea's suggested slide count.
• Video ideas open as a vertical 1080×1920 video (Pro and Ultra).

Need a different size? Copy the brief from the detail panel and start it fresh in the AI Designer with the size you want.

Brand Mind does the heavy lifting

The brief that goes to the AI is about content and direction, not colors and fonts. That's deliberate: your Brand Mind is injected into every AI design automatically, so the idea says what to post and Brand Mind says what it should look and sound like. If you have a Brand Kit (Pro and Ultra), it's applied too.

Photos and video inputs

Some ideas ask for an image or video (a product shot, a clip). You can upload it in the detail panel. When the design opens, attach that photo in the chat if the AI hasn't placed it, or drag it onto the canvas. Text inputs are added to the brief directly.

After it's generated

You're in a normal AI Designer conversation, so everything from there works as usual:

• Refine in chat. "Bigger headline", "use the customer photo as the hero", "make it a 5-slide carousel instead."
• Keep or discard the AI's changes from the second message on.
• Edit by hand in the editor.
• Export when it's ready.

The idea's copy framework and CTA are in the detail panel whenever you need them for the caption.

Batching a week of content

A quick workflow that works well:

1. Generate a batch of ideas (5 credits) and delete the ones that don't fit.
2. Filter by pillar and pick one idea per pillar for the week.
3. Create Visual on each. Each opens in its own editor tab-worth of work; finish one, come back, do the next.
4. Toggle Show created off to hide what's done and see what's left.

Tips

• Fill required inputs with real material. The AI designs around your actual quote or number, and it shows.
• Read the visual direction before you click. If you disagree with it, add a line in the chat once the design opens rather than fighting it after.
• Ideas you've created stay in your Visuals library like any other design, so the Post Ideas page is your backlog, not your archive.