Philip Yao
It's kaggle for prompt engineering.
Core trick
Scores user-submitted prompts against hidden test cases via OpenAI or Anthropic APIs and ranks results on a leaderboard.
Tags
Description
It's kaggle for prompt engineering.
Creating prompts for multiagent systems is too painful. I built a platform to crowdsource this in a competition format
README excerpt
# OptoPrompt - Prompt Competition Platform A Kaggle-style platform for crowdsourced prompt engineering competitions. ## Features - **Competition Listing**: Browse all available prompt engineering competitions - **Competition Details**: View detailed information, test cases, and leaderboards - **Multi-step Competition Creation**: Create competitions with custom test cases and prizes - **Prompt Submission**: Submit prompts and get automated evaluation - **User Profiles**: Track your submissions and created competitions - **Automated Evaluation**: Prompts are evaluated against test cases using OpenAI or Anthropic APIs ## Tech Stack - **Frontend**: Next.js 14 (App Router), React, TypeScript, Tailwind CSS - **UI Components**: shadcn/ui - **Backend**: Next.js API Routes - **Database**: PostgreSQL (Neon) with Prisma ORM - **Authentication**: Firebase Auth - **Storage**: Firebase Storage (training/validation data) - **Encryption**: AES-256-GCM for API keys - **LLM Integration**: OpenAI, Anthropic, Google AI (via Vercel AI SDK) ## Getting Started ### Prerequisites - Node.js 18+ - PostgreSQL database - OpenAI API key (for GPT models) - Anthropic API key (for Claude models) ### Installation 1. Clone the repository: ```bash git clone <repository-url> cd Optoprompt ``` 2. Install dependencies: ```bash npm install ``` 3. Set up environment variables: ```bash cp .env.example .env ``` Edit `.env.local` and add: - `DATABASE_URL`: Your PostgreSQL connection string - `FIREBASE_SERV