This repository automatically generates new single-file apps (either HTML or Python) every day using the Perplexity AI API, powered by GitHub Actions automation.
Each run creates a new prompt, builds a working app, and commits the results back to the repository for version tracking.

The system works in two main stages:

1. Prompt Generation – A creative, structured product specification is generated using Perplexity.
2. App Generation – Based on the prompt, Perplexity builds a complete single-file application that can be an HTML app (with inline CSS and JavaScript) or a pure Python script.

All results are automatically saved in organized folders under dist/:

• dist/prompts/ – versioned prompt text files describing each generated app
• dist/apps/ – versioned generated applications (.html or .py)
• dist/latest/ – the most recent prompt and app for quick access
• dist/log.txt – log of all runs with timestamps and statuses

automation/ │ ├── generate.py → Main script that handles prompt generation and app creation │ .github/ │ └── workflows/ └── daily-perplexity.yml → GitHub Actions workflow that automates daily code generation │ dist/ │ ├── apps/ → Generated applications (HTML or Python) ├── prompts/ → Generated prompt specifications ├── latest/ → Latest generated prompt and app for quick access └── log.txt → Log file containing timestamps and status of each run

• API Endpoint: https://api.perplexity.ai/chat/completions
• Model: sonar (can be overridden via workflow environment variable)
• Automation: GitHub Actions runs once per day at 13:00 UTC (≈ 9 AM US Eastern)
• Retries: The workflow automatically retries failed API calls up to three times, with exponential backoff.
• Timeouts: Each API call allows up to 420 seconds for a response.
• Logging: Every success or failure is appended to dist/log.txt for easy debugging.

1. GitHub Action triggers daily (or manually via Run workflow).
2. The Python script generate.py runs and:
   • Generates a creative prompt for a new single-file app.
   • Builds the complete app (HTML or Python) based on that prompt.
3. Generated files are versioned (no overwrites). If a filename already exists, it’s automatically saved as -2, -3, etc.
4. All new files are committed and pushed back to the repository with the message “automation using Perplexity AI.”

1. Clone or fork this repository.
2. In your repository settings, go to Settings → Secrets and variables → Actions → New repository secret.
3. Add a secret named PPLX_API_KEY with your Perplexity API key.
4. Commit your workflow YAML file (.github/workflows/daily-perplexity.yml).
5. Run the workflow manually once to verify the setup.