PromptFixer

PromptFixer is a service designed to help users improve their AI prompts by collecting feedback and issues, then delivering enhanced prompts. The platform offers two primary methods for users to submit their prompt fixing requests: a web form and email submission. Users can fill out a form with their existing prompt, expected outcome, and a description of the issue, or they can email their prompt details, which are automatically parsed using Gemini AI.

All submissions are stored in a SQLite database and can be accessed through an admin panel. Administrators can view all prompt requests, create improved versions of user prompts, and send the fixed prompts back to users via email. The project structure includes backend server-side code, frontend client-side code, and shared types and utilities.

The backend is built using the Val Town Platform, Hono API framework, and SQLite for database operations. The frontend consists of a public HTML form for submissions and an admin dashboard. The platform also utilizes OpenAI for email processing, TailwindCSS for styling, and JWT for authentication.

Key features of PromptFixer include a public HTML form that collects user details, email processing to extract prompt information, an admin panel for managing requests, and email notifications to keep users and admins informed. The API endpoints include health checks, form submissions, email processing, admin authentication, and request management.

PromptFixer aims to eliminate the guesswork and frustration often associated with working with large language models (LLMs), chatbots, or retrieval systems. By providing a structured and efficient way to improve prompts, it helps users achieve better results without the need for extensive rewrites or trial and error.

The service is free to use, allowing users to fix up to 5 prompts per day without requiring a signup.