Show HN: A GPT Image 1.5 UI for developers tired of rewriting prompts (gpt2image.io)

Hey HN,

  After burning through dozens of credits rephrasing the same prompt in different AI image tools, I built gpt2image.io - a conversational interface for GPT Image 1.5.

  **The core idea:**
  Instead of prompt archaeology ("was it 'modern minimalist' or 'contemporary minimal' that worked?"), you iterate through dialogue:
  - "Change the background to midnight blue"
  - "Make the text bolder"
  - "Keep everything but replace the laptop"

  Context persists across turns. Style consistency is maintained. It's how humans actually work.

  **Built for:**
  - Developers mocking up UIs
  - Designers exploring concepts rapidly
  - Product teams creating marketing assets
  - Anyone who needs images that evolve with ideas

  **Tech highlights:**
  - Built on Next.js 16 + React 19 (using React Compiler for optimization)
  - Turbopack for stupid-fast local dev
  - Multi-language support out of the box
  - Clean API for integrations

  **What it's not:**
  - Not trying to replace Photoshop
  - Not a magic solution for everything
  - Still bound by underlying model capabilities

  Try it: https://www.gpt2image.io (first 2 images free)

  Looking for feedback on:
  - What editing patterns break down
  - What controls matter most in production workflows
  - API features you'd actually use

  Contact: contact@gpt2image.io