Looking for advice from AI image generation experts 👋
Looking for advice from AI image generation experts 👋
I’m building JewelViz, an AI jewellery photoshoot platform for jewellery businesses.
We’re currently using Nano Banana 2, but I’m exploring whether there’s a more consistent model for jewellery image generation.
The biggest challenge is jewellery accuracy.
I’m looking for a model that can:
- Keep the original jewellery design unchanged
- Preserve stone count, shape, colour and placement
- Maintain metal details and proportions
- Generate consistent results across different models, poses and settings
- Work reliably across different jewellery types
For us, generating a beautiful image is not enough. The jewellery itself needs to remain accurate.
If you’ve worked with image generation/editing models or APIs for product photography:
Which model would you recommend for this use case, and why?
Would love to hear your real-world experience, especially with jewellery or other highly detailed products.
One thought that might help with jewellery accuracy is using very controlled camera settings during the reference capture stage. Consistent lighting, fixed angles, and stable exposure can reduce the tiny reflections and light fractioning that confuse models. Even if the final output is AI‑generated, having uniform source material often makes the model behave more reliably. Just an idea — hope it helps a bit!
Hey Suhail, JewelViz sounds like an awesome project. Jewellery accuracy is definitely one of the toughest challenges in AI imaging — tiny reflections, metal textures, and stone details can throw models off. I’m not an expert in jewellery specifically, but I’ve seen consistency improve a lot when models are trained with tightly‑controlled lighting and angle references. Wishing you the best with the platform — excited to see where you take it.