Skip to main content

Best Practices for Creating a Text Prompt

Best practices and tips on text prompts for 3D generation

Updated this week

This guide is designed for experienced users who want to push the boundaries of Meshy AI and achieve highly specific, creative, and polished 3D models. Beyond the basics, advanced prompt engineering involves using a structured approach to leverage complex keywords, negative prompts, and detailed artistic styles.

The In's and Out's of an Advanced Prompt

A powerful Meshy prompt is more than just a list of keywords. It can be broken down into three key components:

  1. Subject: The core object or character you want to create (e.g., "a robot," "a knight," "a futuristic car").

  2. Modifiers: Words that describe the subject's properties, materials, colors, and specific details. This is where you add complexity and nuance.

  3. Styles: Keywords that define the overall aesthetic, artistic genre, or visual characteristics of the model.

Example of a structured prompt: A fearsome dragon, (Subject) with polished obsidian scales, glowing red eyes, and intricate bone armor, (Modifiers) in the style of a dark fantasy video game concept art, highly detailed, cinematic lighting. (Styles)

Mastering Advanced Modifiers and Keywords

Meshy responds to a vast vocabulary of descriptive terms. Using these effectively can dramatically change your results.

Common Advanced Modifiers:

  • Materials: polished metal, aged wood, rusty iron, iridescent glass, carved stone, luminescent crystal, cracked leather

  • Details & Textures: intricate carvings, fine filigree, smooth surface, rough texture, futuristic circuitry, glowing runes

  • Lighting & Atmosphere: cinematic lighting, dramatic shadows, soft ambient light, volumetric fog, ethereal glow

  • Poses & Stances: T-pose, A-pose, dynamic action pose, in motion, sitting, standing (useful for character models)

  • Quality & Realism: highly detailed, ultra-realistic, photo-realistic, masterpiece, 8k resolution, Unreal Engine, Octane render

Example: A steampunk-style owl, with intricate clockwork wings made of brass and copper, glowing gears, and a monocle on one eye. Cinematic lighting, smooth metal texture, ultra-detailed.

Exploring Specific Artistic Styles

Meshy can generate models in a wide range of artistic styles. By combining keywords, you can create unique aesthetics that are tailored to your project.

Examples of advanced style combinations:

  • Cyberpunk Neon: cyberpunk, neon glow, holographic, sci-fi, dark atmosphere, Japanese street style

  • Fantasy Diorama: diorama, miniature scale, highly detailed, hand-painted, clay-sculpted, whimsical, fantasy art

  • Ancient Ruins: ancient ruins, overgrown with vines, cracked stone, mossy, mysterious, cinematic, volumetric lighting

  • Anime Cell Shading: cell shaded, anime art style, sharp lines, vibrant colors, clean edges, stylized

  • Low Poly Aesthetic: low poly, simple shapes, vibrant colors, flat shading, geometric, retro video game style

Example: A magical tree, with leaves made of sapphire crystals, glowing runes on its bark, and floating ethereal wisps around it. Low poly style, vibrant colors, isometric perspective.

The Iterative Process

Advanced prompt engineering is a process of refinement. Start with a simple prompt and then:

  1. Generate: Create a few variations to see what works.

  2. Refine: Analyze the results. What did you like? What did you dislike?

  3. Adjust: Add more descriptive keywords, update your negative prompts, or try a different style combination.

  4. Repeat: Continue this process until you achieve the desired outcome.

By mastering these techniques, you'll be able to unlock the full creative potential of Meshy AI, turning your most imaginative ideas into stunning 3D models with precision and control.

More info on Text to 3D model generation can be found here.

Did this answer your question?