1
Open S.I.M.P.
Go to simpai.netlify.app. No account, login, or installation required — the app runs entirely in your browser.
2
Paste your raw prompt
Click the input field and paste the prompt you want to analyze. This should be the prompt exactly as you’d send it to an AI model — unpolished, unedited.
3
Choose your mode
Select Be Nice for constructive, senior-engineer-style feedback, or Roast Me if you want the same analysis delivered with considerably less diplomacy. Both modes produce the same structured output; only the tone differs.
4
Run the analysis
Click Run S.I.M.P. (or Roast This Prompt in Roast Me mode). S.I.M.P. sends your prompt to the backend, which analyzes it and returns your results.
5
Review your risk score and issues
Your result includes a hallucination risk score from 1 to 10, a risk label (Low, Moderate, High, or Critical), and a list of specific issues detected in your prompt. Read through the issues to understand what made your original prompt risky.
6
Copy your sanitized prompt and take it to an AI model
Review the sanitized prompt and the change log explaining what was fixed. When you’re ready, use the one-click handoff to open the sanitized prompt directly in ChatGPT, Claude, or Gemini — or copy it manually.
Example
Original prompt (before S.I.M.P.)
Sanitized prompt
S.I.M.P. cleans prompts — it doesn’t answer them. Always take your sanitized prompt to an AI model.