Picva
Search, filter, and reuse high-quality prompts for multiple AI image models.
Search: Tech
Prompt share: Blueprint 💬Prompt: A blueprint schematic of a [subject], drawn in the style of early 20th-century industrial patents. Rendered in crisp blue ink with white technical lines, featuring exploded views, angular labels, and stamped diagram codes. Check ATLs for inspiration ✨
A handcrafted illustration that simulates traditional embroidery using colorful threads on linen fabric. All elements are “stitched” with visible yarn textures, using techniques like satin stitch, backstitch, and French knots. Raised contours and directional thread flow create a tactile, cozy appearance. The background is made of woven linen, with gentle pastel or folk-inspired colors. The composition is friendly and magical, evoking a storybook charm. Include decorative details such as flowers,
Transform this image into a stylized mechanical design sheet inspired by Japanese anime mecha blueprints.\nRender the subject with precise, clean linework and semi-realistic cel-shading, as if it were a technical illustration from an anime artbook.\nBreak down the design into parts (e.g. head, hand, accessory, full body) and arrange them like a reference sheet, with different angles and views.\nInclude soft handwritten notes or Japanese characters scattered like concept annotations.\nUse muted d
Tip 2: Use Text Rendering Parameters Nano Banana Pro's secret weapon? Perfect text rendering in images. But only if you tell it exactly what text, where, and in what style. Template Prompt: { "scene": "[your image description]", "text_elements": [ { "text": "[exact text to render]", "position": "[where in image]", "font_style": "[typography style]", "language": "[if not English]" } ] } -- Example Prompt: { "scene": "minimalist tech poster with gradient background", "text_elements": [ { "text": "
{ "style": "layered paper-cut illustration, papercraft diorama, handcrafted aesthetic", "technique": { "layering": "multiple stacked paper layers with soft drop shadows between each layer", "depth": "5–7 visi