A recipe book for GPT-4o’s image mode

What it does

This repository is a curated gallery of GPT-4o image generation prompts. Each entry bundles the full prompt text, a generated example image, and author attribution scraped from X and WeChat. Everything is packaged into a single Prompts.html file that acts as a browsable dictionary.

The interesting bit

The collection is aggressively practical: no theory, just working prompts with proven outputs. It mixes English and Chinese entries, reflecting where the actual GPT-4o prompt innovation is happening right now—mostly on social media rather than in research labs.

Key highlights

• 76 prompts covering styles from 3D emoji to blueprint schematics to “fluffy logos”
• Every entry includes the exact prompt text, a generated example image, and the original creator’s social handle
• Bilingual content (English and Chinese) sourced from active community creators
• Distributed as a single HTML file, making it browsable without a build step
• MIT licensed, so you can reuse the prompts without legal hand-wringing

Caveats

• Several prompt descriptions are truncated in the README, so you may need to open Prompts.html for the full text
• Fundamentally a curated content list, not a tool or framework; expect copy-paste, not automation
• Heavy reliance on externally hosted images (cdn.imgedify.com) means the gallery could degrade if those links rot

Verdict

Grab this if you generate images with GPT-4o and want to reverse-engineer working prompts by example. Skip it if you are looking for a programmatic API, prompt chaining logic, or original research—it is a dictionary, not a laboratory.