Seed
A seed is an integer that initializes the random number generator inside an image, video, or audio model, making generation reproducible.
Pass the same seed + same prompt + same model + same sampler and you get the same output, every time. This is essential for iterating on a prompt — fix the seed, change one word, and you can attribute the visual change to that word alone. Most diffusion APIs (Stable Diffusion, Flux, Midjourney via --seed, Replicate) expose the seed. LLM APIs increasingly support a `seed` parameter too for partially-reproducible text output (OpenAI, Anthropic with extended thinking off). Sharing a seed alongside a prompt makes your result reproducible to others.
When to use seed
- A/B testing prompt edits — fix the seed.
- Generating four variants — change only the seed.
- Sharing reproducible results in tutorials.
Common mistakes
- Reporting a successful prompt without its seed — readers can't reproduce.
- Assuming seed reproducibility carries across model versions — it usually doesn't.
FAQ
What is seed?
A seed is an integer that initializes the random number generator inside an image, video, or audio model, making generation reproducible.
When should I use seed?
A/B testing prompt edits — fix the seed. Generating four variants — change only the seed. Sharing reproducible results in tutorials.
What are the most common mistakes with seed?
Reporting a successful prompt without its seed — readers can't reproduce. Assuming seed reproducibility carries across model versions — it usually doesn't.
Related terms
- Diffusion model — A diffusion model is a generative neural network that creates images, video, or audio by iteratively denoising random noise toward a learned target distribution.
- CFG scale (classifier-free guidance) — CFG scale controls how strongly a diffusion image model follows its text prompt — higher values stick closer to the prompt, lower values explore more.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/seed.md.