ChatML Formatter
Convert a conversation into ChatML, Llama 3, Mistral, Gemma or Alpaca prompt templates.
Build detailed image prompts for Midjourney, Stable Diffusion and DALL-E.
Be concrete. What is in the picture, and what is it doing?
Paste this into the separate negative prompt field, not the main one.
Terms near the front of a prompt carry more weight in most image models, so the subject leads, the setting follows, and the visual treatment comes last. If something is not coming out right, move it earlier before you add more words about it.
The AI Image Prompt Generator builds detailed image prompts in the order image models actually respond to: subject first, then action and setting, then the visual treatment — art style, lighting, colour, lens and composition. Output is tuned separately for Midjourney, Stable Diffusion and DALL-E.
Image models respond to detail in a fairly predictable order: subject first, then what it is doing and where, then the visual treatment — art style, lighting, colour, lens and composition. This builder lays those out as fields with real options behind them, so you get a prompt with the descriptive weight in the right places instead of a bare noun and a hope.
Output is tuned per target. Midjourney gets its double-dash parameters for aspect ratio, stylize and version; Stable Diffusion gets a separate negative prompt, since that is where its quality usually comes from; DALL-E and the chat-based generators get a plain descriptive sentence, because they read natural language rather than comma-separated tags. Nothing is generated here — this writes the prompt, and you paste it wherever you make images.
Fields follow the order image models respond to, so you get a considered prompt rather than a bare noun and a hope.
Double-dash parameters for aspect ratio, stylize and version are appended in the correct syntax.
A separate negative prompt is generated, since that is where Stable Diffusion quality usually comes from.
Chat-based generators get a plain descriptive sentence, because they read language rather than comma-separated tags.
Art style, lighting, colour, lens and composition come with curated choices instead of an empty text box.
Midjourney, Stable Diffusion or a natural-language model such as DALL-E. The output format changes to match.
Say what is in the picture and what it is doing. Be concrete: "a fox curled asleep on a windowsill" beats "an animal".
Pick style, lighting, colour, camera and composition, or leave any of them out to give the model more freedom.
Copy the finished prompt, including model parameters and any negative prompt, into your image tool.
A list of things you do not want in the image. Stable Diffusion takes it as a separate field and it does real work there. Midjourney uses the --no parameter, and DALL-E has no equivalent, so the wording is folded into the sentence instead.
Yes. Terms earlier in the prompt carry more weight in most models, which is why this builder puts the subject first and the finishing details last.
Match it to where the image will live: 16:9 for a header or slide, 1:1 for a social post, 9:16 for a story or reel, 3:2 for something that reads as a photograph.
No. It writes the prompt text, in your browser. You take the prompt to Midjourney, Stable Diffusion or whichever tool you use.
What people who use the AI Image Prompt Generator think of it. Reviews come from visitors and are posted as written.
No ratings yet
Yours would be the first.
0 reviews
Newest first
No reviews yet
Be the first to rate the AI Image Prompt Generator. It takes half a minute and no account.
Write the first reviewConvert a conversation into ChatML, Llama 3, Mistral, Gemma or Alpaca prompt templates.
Generate robots.txt rules to block GPTBot, ClaudeBot, CCBot and other AI crawlers.
Estimate how many tokens your text costs before you send it.
Estimate the cost and vector storage of embedding a document corpus for RAG.
Turn question and answer pairs into a JSONL training file for OpenAI, Llama or Gemini.
Write valid tool-calling JSON schemas without hand-editing braces.
Compare two embedding vectors by cosine similarity, dot product and distance.
Turn a rough idea into a structured prompt that gets a usable answer.