Your prompt29 words
a cat wearing a tiny songkok, a neon-lit Kuala Lumpur street at night, photorealistic, close-up portrait, 85mm portrait lens, neon glow, whimsical mood --ar 4:5 --no blurry, text, watermark
Negative prompt:
AI Prompt Generator for Images and Video
Describe your idea once and get 3 ready-to-paste prompts (as chosen, cinematic and artistic), written in the format each app reads best: Midjourney, Stable Diffusion, ChatGPT/DALL·E and Flux for images, Veo, Sora, Kling and Runway for video. Click the chips below or press Surprise me.
Your prompts for Midjourney
Cinematic twist38 words
a cat wearing a tiny songkok, a neon-lit Kuala Lumpur street at night, cinematic film still, close-up portrait, 85mm portrait lens, neon glow, whimsical mood, teal and orange, film grain --ar 4:5 --stylize 150 --no blurry, text, watermark
Negative prompt:
Artistic twist36 words
a cat wearing a tiny songkok, a neon-lit Kuala Lumpur street at night, digital painting, close-up portrait, 85mm portrait lens, neon glow, whimsical mood, pastel palette, intricate textures --ar 4:5 --stylize 400 --no blurry, text, watermark
Negative prompt:
Have a picture instead? Turn an image into a prompt.
Saved prompts
Advertisement
One idea, written for each app
The same cat and the same hawker-stall video, as the generator writes them. Notice where the negative prompt goes and how the video prompts carry the camera move and sound.
| App | Prompt |
|---|---|
| Midjourney | a cat wearing a tiny songkok, a neon-lit Kuala Lumpur street at night, photorealistic, close-up portrait, 85mm portrait lens, neon glow, whimsical mood --ar 4:5 --no blurry, text, watermark |
| Stable Diffusion | a cat wearing a tiny songkok, a neon-lit Kuala Lumpur street at night, photorealistic, close-up portrait, 85mm portrait lens, neon glow, whimsical moodNegative prompt: blurry, text, watermark |
| ChatGPT / DALL·E | A photorealistic photo of a cat wearing a tiny songkok in a neon-lit Kuala Lumpur street at night. Shot as a close-up portrait with an 85mm portrait lens. Neon glow, whimsical mood. Portrait 4:5 format. |
| Flux | A photorealistic photo of a cat wearing a tiny songkok in a neon-lit Kuala Lumpur street at night. Shot as a close-up portrait with an 85mm portrait lens. Neon glow, whimsical mood. |
| Google Veo | A medium shot of a chef tossing noodles in a flaming wok in a busy hawker stall at night. Camera: slow dolly in. Neon glow. The character says: "Siap!" Audio: sizzling wok, crowd chatter. |
| Runway | Slow dolly in: A medium shot of a chef tossing noodles in a flaming wok in a busy hawker stall at night. Neon glow. |
| Kling | A medium shot of a chef tossing noodles in a flaming wok in a busy hawker stall at night. Camera: slow dolly in. Neon glow.Negative prompt: blurry |
How each app reads a prompt
- Midjourney: short comma phrases, then parameters:
--arfor the shape,--nofor things to leave out,--stylize(0–1000) for how artistic. - Stable Diffusion: comma keywords, most important first, with a separate negative prompt. Only about the first 75 tokens are read.
- ChatGPT/DALL·E, Flux, Gemini: full sentences, subject first. There is no negative prompt; say what you want instead.
- Veo and Sora: one shot with the camera move, plus sound and any spoken line in quotes.
- Runway: the camera move first, then the scene; keep every instruction positive.
- Kling: a descriptive scene in the main box and short, specific negatives in its negative prompt box.
Advertisement
How to use the AI Prompt Generator
- Pick Image or Video and the app you will paste into.
- Type the subject, and optionally the action and setting. Or press Surprise me for a random idea.
- Tap chips for style, lighting, framing, mood and colour (camera movement and speed for video).
- Compare the 3 versions, copy the one you like, and save favourites for later.
- Share the link: it rebuilds exactly the same prompt for anyone who opens it.
Sources
- Midjourney: parameter list
- Runway: AI video prompting guide and Runway Help Center
- Kling AI: negative prompts
- Replicate: using and prompting Veo 3
- fal.ai: how to use Flux
- Hugging Face forums: the 77-token CLIP limit
Note: AI apps change their models often, so results vary between versions. Follow each app's content rules, and do not use prompts to copy a living artist's work or a real person's likeness without permission.
Frequently asked questions
How do I write a good AI image prompt?
Put the subject first, then where it is, then the style, framing, lens, lighting, mood and colours. For example: a cat wearing a tiny songkok, a neon-lit Kuala Lumpur street at night, close-up portrait, 85mm portrait lens, neon glow, whimsical mood. Specific words beat long lists of vague ones such as 'beautiful' or '8k'.
What do --ar, --no and --stylize mean in Midjourney?
They are Midjourney parameters added after the prompt. --ar sets the aspect ratio (--ar 16:9 is wide, --ar 9:16 is a phone screen), --no lists things to leave out (--no text, watermark), and --stylize (0 to 1000) sets how much Midjourney's own artistic style is applied. This tool adds --stylize 150 to the cinematic version and 400 to the artistic one.
Which AI apps support negative prompts?
Stable Diffusion and Kling have a separate negative prompt box, and Midjourney uses --no. ChatGPT/DALL·E, Flux, Runway and Veo have no negative prompt syntax, so describe what you want instead; Runway notes that negative wording can even produce the opposite. The generator puts your negatives in the right place for each app, or leaves them out and tells you.
How long should a Stable Diffusion prompt be?
Stable Diffusion's CLIP text encoder reads 77 tokens, about 75 of them usable, and quietly drops the rest. That is roughly 50 to 60 words. The card shows an approximate token count for Stable Diffusion and warns past 75, so put the most important words first.
How do I write a prompt for AI video like Veo, Sora, Kling or Runway?
Describe one shot: the subject and one action, the setting, the camera movement (slow dolly in, orbit, drone flyover), then style, lighting and mood. For Veo and Sora add sound, and put spoken lines in quotes. For Runway start with the camera move, as in 'Slow dolly in: a chef tossing noodles in a flaming wok', and keep the wording positive.
Should I write prompts in English or Bahasa Melayu?
English. Most image and video models learned mainly from English captions, so they follow English prompts most closely. The style chips on this page are already in English; you can type the subject in Malay, but translating it to English usually gives a more accurate picture.
Is the AI Prompt Generator free, and are my prompts saved?
It is free with no sign-up and runs entirely in your browser; nothing you type is sent to this site. Prompts you save with the Save button stay in your browser's own storage on this device, and each prompt can be shared as a link that rebuilds it.
Share this calculator
Advertisement