How they work
Most current image generators are diffusion models: they start from random noise and remove it step by step, guided by your text, until an image appears. The text is turned into numbers by a language encoder, so clear, concrete words matter.
The main tools
- DALL-E (OpenAI) — DALL-E 3 arrived in 2023 and follows long, detailed prompts closely.
- Midjourney — an independent lab's service, known for a strong artistic style; version 6 came out at the end of 2023.
- Stable Diffusion (Stability AI, 2022) — its weights are public, so it can run on your own computer; SDXL followed in 2023.
- FLUX.1 (Black Forest Labs, 2024) — from researchers behind Stable Diffusion, with some versions available to run locally.
Anatomy of a good prompt
Subject
What is in the picture: "a ceramic teapot on a wooden table".
Style or medium
Photograph, watercolour, 3D render, line drawing.
Composition
Close-up, wide shot, from above, centred, with empty space on the left.
Light and colour
Soft morning light, warm tones, high contrast.
Format
Square, portrait or wide — set the aspect ratio for where the image will be used.
Some tools also accept a negative prompt (what to leave out) and a seed, a number that lets you reproduce an image and change one thing at a time.
Beyond text-to-image
- Image-to-image — start from a sketch or photo and restyle it.
- Inpainting — repaint only a selected area.
- Upscaling — enlarge an image and add detail.
Use them responsibly
- Check each service's terms for commercial use.
- Don't create misleading images of real people or events.
- Say when an image is AI-generated, especially in news or reviews.
Tools for this
Asaiejadoo — everyday calculation and AI guides, part of the ZIBADIS network founded by Masoud Moghaddam in Tehran.
