Hover over any image. Click one button. Get a ready-to-use prompt for Midjourney, Stable Diffusion, FLUX and DALL-E — instantly.
Free • Open source • MIT license • No account required
No complex setup. No API knowledge required.
The 🔮 button appears automatically on any image on any webpage.
One click sends the image to your chosen AI model for analysis.
Get a detailed prompt with positive/negative tags, style, lighting and composition.
Works on all websites — images, artworks, photos, screenshots. Even video frames from YouTube and Vimeo.
Use Ollama, LM Studio or Jan on your own hardware. Fully offline, completely private.
Images are automatically resized to 1024px and compressed before sending — saves bandwidth and memory.
Waiting too long? Hit Cancel in the side panel. Smart timeouts: 120s local, 60s cloud.
All generated prompts are saved in the side panel. Search, copy or delete them anytime.
Generate prompts in English, Russian or Chinese — choose in Options.
No backend server. Your API key is stored only in your browser. Nothing is collected.
Override the default analysis instructions with your own prompt in Options.
Connect any OpenAI-compatible provider
300+ models, free tier available
CloudExtremely fast inference
Cloudlocalhost:11434/v1 — no key
Locallocalhost:1234/v1 — no key
Locallocalhost:1337/v1 — no key
Local