Under the hood.
A clear, honest explanation of how TextToImg turns text into images.
Step 1: You Type a Prompt
You describe what you want in plain language. The more specific you are — mentioning lighting, mood, style, and composition — the better the result. Use our Enhance button and GPT-4o will automatically improve your prompt.
Step 2: We Send It to the AI Model
Your prompt is sent from your browser to our secure PHP backend, which forwards it to the AI model via OpenRouter. Your API key is never exposed in the browser. The AI model processes your prompt using a diffusion or transformer-based architecture to produce an image that matches your description.
Step 3: The Image Is Returned
The AI model returns a URL pointing to your generated image. We pass this URL to your browser so you can view and download it. We do not permanently store your generated images on our servers.
How the Rate Limit Works
Each device gets 5 free image generations per day. This is tracked server-side by your IP address (including Cloudflare real-IP detection). It cannot be bypassed by clearing cookies, using incognito mode, or switching browsers. The counter resets at midnight UTC.
Which Model Should I Use?
- Flux 1.1 Pro — Best for photorealistic images, portraits, and detailed scenes
- DALL·E 3 — Best for illustrations, concept art, and complex multi-element prompts
- Stable Diffusion XL — Great all-rounder, excellent for anime and stylized art
- Flux Schnell — Fastest option, ideal for rapid iteration