How Refunds Work
Learn how refunds work—submit feedback, get approved, and receive credits back. Your feedback improves our AI models.
We want you to be satisfied with every generation. When the AI doesn't deliver what you expected, we've made it simple to request a refund.
How the Refund Process Works
1. Submit Your Feedback
When you're not happy with a generation, click the refund button on the image and tell us what went wrong. Be specific—the more detail you provide, the faster we can review your request.
Examples of helpful feedback:
- "The sofa legs are distorted and don't match the original"
- "The background has visible artifacts near the product edges"
- "The perspective change made the proportions look unrealistic"
2. We Review Your Request
Our team reviews each refund request. We look at the original image, the generated result, and your feedback to determine if a refund is appropriate.
Most requests are reviewed within 24 hours.
3. You Receive Your Credits
If approved, the credits used for that generation are returned to your account. You'll see them reflected in your balance immediately after approval.
What Qualifies for a Refund
Refunds are typically approved when:
- The output has visible quality issues (artifacts, distortions, blurring)
- The AI misunderstood the request significantly
- Technical errors affected the generation
- The result is clearly unusable for its intended purpose
How Your Feedback Helps
Your refund feedback is valuable to us beyond just the refund itself.
Better Training Guides
Common issues often point to gaps in our documentation. If many users encounter similar problems, we update our guides to help everyone get better results from the start. The chat-to-edit and edit a specific area guides cover the fixes that most often turn an unusable generation into a publishable one, and the uncanny valley of furniture images explains the rendering problems behind most refund requests.
For the full reference-to-review process, see how AI furniture photography works.
Product Development
Feedback patterns inform our roadmap. If certain use cases consistently underperform, we prioritize improvements to those tools and workflows.
If you are tracking recurring failures during a trial, the AI image-platform evaluation checklist helps organize the quality and rework checks.
Tips for Successful Refund Requests
- Be specific about the problem — "It looks bad" is harder to evaluate than "The wood grain texture is completely wrong"
- Submit promptly — Request refunds soon after generation while the context is fresh
- One request per generation — Each image can only have one refund request
Frequently asked questions
How do I request a refund for a generation I'm not happy with?
Click the refund button on the image and describe what went wrong. Specific feedback speeds up the review: say that the sofa legs are distorted and don't match the original, or that the background shows artifacts near the product edges, rather than that it looks bad. Submit soon after the generation while the context is fresh, and note that each image can carry only one refund request.
How long does a refund review take?
The team reviews every request, comparing the original image, the generated result and your feedback to decide whether a refund is appropriate. Most requests are reviewed within 24 hours. If approved, the credits used for that generation go back to your account and appear in your balance immediately after approval.
What kinds of problems qualify for a refund?
Refunds are typically approved when the output has visible quality issues such as artifacts, distortions or blurring, when the AI significantly misunderstood the request, when a technical error affected the generation, or when the result is clearly unusable for its intended purpose. The more precisely the feedback describes the fault, the easier it is to evaluate.
What happens to my refund feedback after the review?
It feeds two things beyond the refund itself. Common issues point to gaps in the documentation, so when many users hit the same problem the guides get updated to help everyone get better results from the start. Feedback patterns also inform the roadmap: if certain use cases consistently underperform, improvements to those tools and workflows get prioritized.
Questions about a specific refund? Contact our support team and we'll help you out.
