This repository offers a curated collection of over 1,500 trending AI image prompts sourced from X (formerly Twitter), ranked by engagement (likes). It aims to provide users with high-quality, proven prompts compatible with various AI image generation tools like NanoBanana Pro, GPT Image, and Midjourney. A key benefit is the included system prompt, designed to enhance prompt engineering by transforming simple inputs into detailed descriptions following specific best practices, making it easier for users to generate superior AI art.

How It Works

The project curates prompts based on their popularity (likes) on X, ensuring relevance and quality. It provides a structured prompts.json file for direct download. Additionally, a sophisticated system prompt is offered, which guides AI models by replacing vague terms with precise parameters (e.g., "90mm, f/1.8" instead of "professional looking"), translating feelings into technical terms (e.g., "Kodak Vision3 500T" for "cinematic vibe"), utilizing negative constraints, and encouraging multi-sensory descriptions. This system prompt can be integrated into workflows via tools like n8n, Gemini, or GPTs.

Quick Start & Requirements

Integration with AI tools is facilitated through npx meigen init commands:

• npx meigen init cursor (for Cursor)
• npx meigen init vscode (for VS Code / GitHub Copilot)
• npx meigen init claude (for Claude Code)

This setup allows searching the prompt library directly from code editors without API keys.

Highlighted Details

• Over 1,500 curated AI image prompts, with weekly updates.
• Prompts ranked by engagement (likes).
• Compatible with NanoBanana Pro, GPT Image 1.5, and more.
• Integrated into MeiGen MCP for seamless searching and generation within coding tools.

Maintenance & Community

The project is actively maintained, with a changelog detailing recent prompt additions and improvements. Users are encouraged to open issues for any spotted problems or low-quality content.

Licensing & Compatibility

The repository README does not explicitly state a software license. This lack of clear licensing information is a significant adoption blocker, particularly for commercial use or integration into closed-source projects.

Limitations & Caveats

The raw data scraped from X may contain advertisements or lower-quality content, though curation is ongoing. Due to the ranking method (likes), the prompt collection is biased towards prompts generating "attractive women," even after filtering.