Promptixa

The Problem

AI users, developers, and creators frequently encounter friction when discovering, tailoring, and testing prompts across disparate LLM platforms. Standard prompt repositories lack real-time parameter customization, direct AI model integrations, and structured community collaboration.

The Solution

Architected and deployed Promptixa, a high-performance Flutter mobile and web community platform. Engineered a dynamic variable-parsing parser that automatically converts template placeholders (<topic>, <tone>, <context>) into interactive UI inputs, allowing users to customize and directly launch final prompts into ChatGPT and Gemini with a single tap.

Engineering Highlights & Architecture

• Feature-driven layered Flutter architecture with clean domain separation
• Riverpod AsyncNotifier state management with generated functional providers
• Dynamic prompt customization engine parsing placeholder tags (<topic>, <tone>, <style>)
• Direct AI workflow integration launching ChatGPT and Google Gemini with customized prompts
• Firestore cursor pagination and fallback queries for high-performance discovery feeds
• Google Authentication with transactional username reservation and collision handling
• Community prompt submission workflow with validation and Cloudinary media uploads
• AdMob monetization engine with Native-to-Banner fallback and interstitial frequency control
• Hive local caching for staggered masonry feed layouts to eliminate visual layout shifts
• Automated test suite with 53 passing unit/widget tests and 0 Flutter analyzer warnings

Technical Challenges

Constructing an elastic template regex parser that dynamically binds user variables without UI stutter; executing atomic username reservations with Firestore transactions during Google Sign-In to prevent race conditions; and maintaining a smooth 60 FPS masonry feed by caching aspect ratios locally in Hive to prevent image layout shifts.

Project Outcome & Production Quality

Successfully launched on Google Play and Web with a robust layered architecture, resilient AdMob monetization (Native -> Banner fallback), comprehensive Firestore security rules, and 53 automated unit/widget tests with zero static analyzer warnings.