AI Prompt Refiner

What It Does

• Takes an original prompt from the user.
• Lets the user choose a refinement mode, tone, and use case.
• Returns three prompt variants: refined, concise, and expert.
• Shows the result pack in a focused modal popup for easier copying.
• Explains what improved in a short list.

Setup

1. Open functions/config.local.php.
2. Add your OpenAI API key to openai_api_key.
3. Keep or change the default model in openai_model. The project starts with gpt-5-mini.
4. Optional: add adsense_client and adsense_sidebar_slot to activate the right sidebar ad area.
5. Optional: instead of the local config file, you can provide OPENAI_API_KEY from the server environment.
6. Load /AiPromptRefiner/ in your browser through MAMP.

How To Use

1. Paste your original prompt into the main textarea.
2. Choose the refinement mode that matches your goal.
3. Optionally add tone, use case, goal, and extra context.
4. Click Refine Prompt.
5. Review the generated output in the popup modal.
6. Copy the version you want, or reopen it with the View Refined Output button.

Folder Structure

• index.php – main UI entry point
• css/style.css – full interface styling
• js/app.js – browser-side interactions and API requests
• functions/lib.php – configuration, validation, OpenAI request logic
• functions/config.local.php – local API key and model settings
• functions/api.php – backend endpoint for prompt refinement
• storage/logs – lightweight debug logs

Security Notes

• The functions and storage folders are protected with .htaccess.
• Requests use a CSRF token.
• A simple file-based rate limit helps reduce abuse without a database.

Troubleshooting

• If you see an API key error, check functions/config.local.php or your environment variables.
• If responses fail, confirm that your PHP setup supports curl.
• If folder protection does not work, make sure Apache allows .htaccess overrides.