Decoder. plain-English AI glossary

Text-to-Image

● Core

Creating images from natural language descriptions—"a cyberpunk city at night" produces a novel picture matching the description.

Think of it like

Commissioning a digital artist who reads your description and delivers a painting in seconds.

Example

DALL-E, Midjourney, Stable Diffusion take text prompts and generate photorealistic or artistic images.

How it actually works

Relies on text encoder (CLIP or BERT-style) to convert prompt to embeddings, then diffusion or GAN uses embeddings to guide generation. Model must learn mapping between language and pixels. Prompt engineering (specific adjectives, artist names, style hints) dramatically improve results.

For product teams

Unlocks content creation and design workflows without hiring artists.

For engineers

Text encoder + diffusion backbone; conditioning signal injected via cross-attention; noise schedule controls quality-speed tradeoff.

Image Generation — Specialization of.

Read anything AI without the jargon

Look up any term in plain English, or save terms as you read with the free Chrome extension.

Open Decoder Add to Chrome