Nanobanana / ImDream vehicle image & video prompt workflow

Users documented a stepwise vehicle image and video prompt workflow using Nanobanana and ImDream. - Veo3 noted as a product with built-in sound for immersive results. - ImDream free tier lacks latest models; paid tier may yield stronger outputs. - Video generation requires clearer, more precise prompts. - Consistency and a 'world model' are essential for coherent videos. - Workflow: find vehicle image, send to Nanobanana Pro, request white-background studio shot. - Example edits: brass sport rims, widened low stance, JKW plate, Red Bull decals, red calipers, dual wing. - ImDream free users get 80 daily points; used to generate eight start-end frame videos. - User shared full prompts and pitfalls on youmind and links for reference.
Key insights
Practical Nanobanana→ImDream vehicle-edit workflow
Users described a concrete workflow for vehicle image editing: source a vehicle photo, submit it to Nanobanana Pro to produce a high-quality white-background studio shot, then supply targeted editing prompts that specify wheel style, stance, body width, license plate text, decal positioning, caliper color, and wing design. The shared prompts include exact visual and typographic details (e.g., Helvetica 'JKW', black with red accents) and ordering guidance to keep unrelated areas unchanged. Full prompt sets and a pitfalls log were shared on a public youmind link for peer learning.
Video generation demands prompt clarity and world-model consistency
Users emphasized that video generation is more demanding than still images: it requires clearer, more detailed prompts and model features that ensure temporal consistency. The discussion highlights the importance of a 'world model' that recognizes, remembers, and maintains scene elements across frames; lack of such consistency is a primary challenge. The posts note that paid access to newer models or Veo3-like products with audio may improve immersion and output quality.
Creative prompt capabilities include 'time-travel' and coordinate-based scenes
Nanobanana was described as capable of creating images conditioned on coordinates, date, and time—framed by one user as effectively enabling 'time travel' in generated imagery. Practical prompt examples referenced geolocated scene recreation and temporally specific compositions. Users reported leveraging limited daily credits (ImDream free tier: 80 points) to generate short start-end frame videos and iterate on vehicle designs.