Works well for marketing purposes like product promotion, sales announcements, and personal branding. For celebratory content like birthday wishes or party invitations, it's still learning.
Sivi prioritizes giving you control over your designs. That's why it generates designs in layers, allowing for easy customization. This process can take a minute or two.
Sivi is constantly evolving! While some initial designs might not perfectly match your vision, we're always working to improve the model's accuracy.
Specify Colors: Provide names or Hex codes via chat (e.g., "#FF5733" or "Coral").
Preferences: Choose from your saved color preferences.
Let Sivi Decide: Ask Sivi to suggest a suitable color palette.
Coming soon: Theme, backdrop styles, text styles. Let us know your ideas for more.
Identify the purpose: This task in prompt engineering is to identify the purpose of the design from the initial prompt. This ensures the text generator crafts copy that perfectly aligns with your goals. Currently, Sivi understands various purposes like product promotion, brand awareness, and personal branding. But that's not all – creating birthday wishes, showcasing digital products with device mockups, and designing party invitations are on the way.
Write content: Based on the identified purpose, the text generator writes a clear and concise copy using the supported semantic text blocks (title, button, offer, coupon, contact, etc.). Feel free to customize the generated copy or provide your own – Sivi gives you complete control over the final message!
Choose Dimensions: Set your size or let the AI Design Generator choose one from popular options such as Instagram post, product banner, YouTube thumbnail, etc. Sivi supports custom dimensions beyond pre-defined aspect ratios such as 1:1, 2:1, etc.
Add Asset Links: Upload your brand assets (image, vector, or logo) for a cohesive look in one of the following ways:
Provide Asset URL - You can paste your brand asset URL
Generate Assets - Generate one with the image generator, copy the image address, and provide that in the prompt.
Upload assets - Upload an image and submit. Once submitted, right click on the image, copy the image address and provide that in the prompt.
Add colors: Provide brand colors or let the AI design generator choose an appropriate palette for you.
Set Language: Choose the language for your design (Sivi supports 72 global languages!).
Here's a quick walk through of how you can use Sivi AI Designer to generate designs.
Troubleshooting prompts and responses of Sivi AI Designer
Request not submitted: If the AI design generator says the request has been submitted without calling the API, just prompt it to submit the request properly. Example prompts: “Can you submit the request properly?”, “Can you generate the designs?”
Image URL issues: Ensure the image URL for uploaded/generated images is correct. If it throws as error, right-click on the image and copy the address before adding it to the prompt.
Incorrect design links: If you receive dummy design links, prompt Sivi to submit the request properly and provide a valid link.
Insufficient credits: If the API throws an error due to insufficient credits, buy credits and resubmit the request.
I submitted a request, but, I didn’t get the designs.
It might be because of one of the following reasons:
Generation failed (review the design brief json).
Generation request not submitted properly (resubmit)
I asked for three variations, but, it generated only one.
You might not receive all requested variations if some designs fail during generation.
Why does the generation fail sometimes or throws error?
Error might be because of invalid URLs or parameters passed in the design brief JSON. You can report them.
How many credits do I get?
Free users get 10 recurring monthly credits (increasing with paid plans). For influencer or educator programs, contact Sivi support.
Why don’t the designs appear instantly?
Sivi processes the design brief, analyses your content, and then generates the designs from scratch, it takes a minute or two (we are working to reduce the time).
Can it generate carousels and adapt across devices?
Not at the moment. It’s in our roadmap.
