By Lucas Desfossez · Rawrr project operator
Local AI can help designers explore visual directions on their own workstation. The useful question is not just “can it make an image?” but “can I control the result, repeat the workflow and use the output in a real design process?”
Images, videos and 3D
Explore silhouettes, materials, surface treatments and visual directions through image generation. Use video to showcase a product in motion. Experiment with 3D for spatial exploration: this part of Rawrr is still experimental and cannot yet reproduce precise objects. Think exploratory shapes, not finished CAD.
Your visual identity, with LoRAs
If a model is a language, a LoRA is the accent you speak it with. If a model is a steak, a LoRA is the salt and pepper you put on it. The model provides the foundation; the LoRA adds character and can dramatically change the output.
LoRAs let you shape the style of your outputs, explore surface treatments for products and cars, or capture a product’s identity and carry it into new scenes across image workflows. How closely that identity holds depends on the model, training and settings.
Your work stays with you
Keep confidential briefs, client assets and unreleased concepts on your workstation. Build your own library of references, generated assets and LoRAs, with control over what you keep and what you share.