- Be Clear and Specific ✅
Avoid vague prompts. The more detail you provide, the better the response.
Example:
Weak: “Explain marketing.”
Strong “Explain digital marketing strategies for small e-commerce businesses.”
Most high-performing prompts follow a specific structure. You don’t need every element every time, but including them ensures our AI doesn't have to guess:
- Persona: "You are a senior research analyst..."
- Task: "Rewrite this passage in a humorous tone."
- Context: "Revise this section as if it were written for a board of directors that is non-technical."
- Format: "Rewrite as a 5-bullet executive summary."
- Use Negative Constraints at the End 🚫
AI models can sometimes "forget" instructions if they are buried in the middle of a long prompt. Place your "What NOT to do" rules (e.g., "Do not use industry jargon" or "Write in first person") as the very last line of your prompt to ensure they stay top-of-mind for the model.
- Provide "Few-Shot" Examples 🎯
If you want a specific tone or a very particular formatting style, don't just describe it—show it. Provide 1 or 2 examples of a "good" response within your prompt. This technique, called few-shot prompting, significantly reduces the need for back-and-forth editing.
- Keep Instructions Simple 📖
With your book, screenplay, or editing instructions, write in simple sentences without bullet points or outline format (Roman numerals). Simple instructions lead to better output.
- Iterate, Don't Restart 🗣️
If the first response isn't perfect, don't delete it and try a new prompt. Treat it as a conversation. Say: "The tone is too formal, make it punchier," or "Focus more on the financial section." Our LLMs remember the previous context, and iterating is often faster than rewriting the original prompt.
- Provide Examples ✅
Show what you want so the model can match your style or format.
Example: “Write this screenplay in the style of a Stephen King novel.”
- Set Constraints 📋
Limit the response to improve relevance. Examples:
- “Use simple language.”
- “Keep it under 150 words.”
- “Avoid technical jargon.”
- Combine Multiple Instructions 👯♂️
You can stack requirements in one prompt, but keep the directions clear and simple. Example:
“Explain AI in simple terms, include 3 real-world examples, and keep it under 200 words.”
- Ask for Alternatives or Options 🧠
Get more creative or useful outputs by requesting variations. Examples:
- “Give me 5 chapter title ideas.”
- “List 3 different approaches to solve this problem.”
- Check spelling and grammar; illicit language or phrasing ❌
Spelling and grammar errors can produce unwanted outcomes in your prompts, so check these carefully.
Squibler’s language models currently restrict cursing and words/phrases that the AI perceives to be violent, erotic, or racist. You can read more about these limitations in our Terms of Service located here: