SiloPrompts - Personal AI Prompt Database

In the GenAI era, every ask, every query, and every search is a prompt.

The ones that actually work — refined through trial and error, containing your proprietary workflows — those are valuable. And most people store them in scattered notes, buried chat histories, or nowhere at all.

Cloud prompt managers exist, but they require uploading your prompts to someone else's server. Your sensitive instructions and strategies sitting on a platform you don't control.

SiloPrompts is a private, self-hosted prompt library. Open source. No cloud. No accounts. No data leaving your machine.

A lightweight, self-hosted web application for storing, searching, and managing your AI prompts across different platforms (ChatGPT, Claude, Perplexity, Gemini, etc.).

Features

• ✏️ Full CRUD — Create, edit, delete prompts from the web UI
• 🔍 Full-text search with real-time results (⌘K shortcut)
• 📋 One-click copy to clipboard
• 🏷️ Tags — Cross-category organization with multi-select filtering
• ⭐ Section-level favorites — Bookmark specific prompts, not just files
• 🔄 Sort — By name, date modified, or size
• 📥 Import/Export — Upload and download .md prompt files
• 🌗 Dark/Light mode with glassmorphism theme
• 🐳 Docker-ready with Kubernetes/Helm support
• 🔐 Secure — Path traversal protection, XSS escaping
• 💾 100% local — Markdown files, git-friendly, no database

Quick Start

pip (Simplest)

pip install siloprompts
siloprompts

Access at http://localhost:5000

Prompts are stored in ~/.siloprompts/prompts/ by default.

# Custom port and prompts directory
siloprompts --port 5002 --prompts /path/to/prompts

# Open browser automatically
siloprompts --open

# Show version
siloprompts --version

Docker

docker run -d -p 5002:5000 bdharavathu/siloprompts

Access at http://localhost:5002

Optional: Mount your own prompts folder for persistence

docker run -d -p 5002:5000 -v ./prompts:/app/prompts bdharavathu/siloprompts

Docker Compose

git clone https://github.com/bdharavathu/siloprompts.git
cd siloprompts
docker-compose up -d

# Access at http://localhost:5002

From Source

git clone https://github.com/bdharavathu/siloprompts.git
cd siloprompts
pip install -e ".[dev,server]"
siloprompts --port 5000 --prompts ./prompts --open

Usage

1. Browse prompts by category in the sidebar or filter by tags
2. Search using keywords — press ⌘K to focus search instantly
3. Click a prompt card → view sections → Copy with one click
4. Favorite individual sections with ☆ for quick access later
5. Create new prompts with "+ New Prompt" or Import existing .md files
6. Edit prompts — add, remove, reorder sections, manage tags
7. Export any prompt as a .md file to share or backup
8. Sort prompts by name, date modified, or size
9. Paste into ChatGPT, Claude, Gemini, Perplexity, or any AI platform

Adding Prompts

Via Web UI

Click "+ New Prompt", fill in the title, category, tags, and sections.

Via Import

Click "↓ Import", select a .md file, choose a category, and upload.

Via File System

Create a .md file in prompts/<category>/:

# My Custom Prompts
Tags: python, debugging

## Prompt Name

​```
Your prompt text here with [PLACEHOLDERS] for variables.
​```

## Another Prompt

​```
Another prompt...
​```

API

GET    /api/prompts                      # List all prompts
GET    /api/prompts/<path>               # Get prompt details
POST   /api/prompts                      # Create prompt
PUT    /api/prompts/<path>               # Update prompt
DELETE /api/prompts/<path>               # Delete prompt
GET    /api/prompts/<path>/download      # Download .md file
POST   /api/prompts/import              # Import .md file
GET    /api/categories                   # List categories
GET    /api/tags                         # List tags with counts
GET    /api/search?q=<query>             # Search prompts
GET    /health                           # Health check

Kubernetes Deployment

helm install siloprompts ./helm/siloprompts
helm upgrade siloprompts ./helm/siloprompts

See helm/siloprompts/values.yaml for configuration: replicas, resources, ingress, TLS, persistent volumes.

Configuration

Environment Variables

Variable	Default	Description
PROMPTS_DIR	~/.siloprompts/prompts (pip) / /app/prompts (Docker)	Prompts storage directory
DATA_DIR	~/.siloprompts/data (pip) / /app/data (Docker)	Application data directory
FLASK_ENV	production	Flask environment
SECRET_KEY	dev-key-change-in-production	Flask secret key
PORT	5000	Server port

CLI Options

siloprompts [OPTIONS]

  --port PORT       Port to listen on (default: 5000)
  --host HOST       Host to bind to (default: 127.0.0.1)
  --prompts PATH    Prompts directory path
  --open            Open browser on start
  --version         Show version and exit

Project Structure

siloprompts/
├── src/siloprompts/       # Python package
│   ├── __init__.py        # Package entry point, version
│   ├── manager.py         # PromptManager class
│   ├── web.py             # Flask app factory and API routes
│   ├── cli.py             # CLI entry point
│   ├── __main__.py        # python -m siloprompts support
│   ├── templates/         # HTML templates
│   └── static/            # CSS, JS, images
├── prompts/               # Default prompt library
├── tests/                 # Test suite
├── helm/                  # Kubernetes Helm chart
├── pyproject.toml         # Package metadata
├── Dockerfile             # Container build
└── docker-compose.yml     # Local Docker setup

Technology Stack

• Backend: Python 3.11, Flask 3.0, Gunicorn
• Frontend: Vanilla JavaScript, HTML5, CSS3
• Storage: Markdown files (no database)
• Container: Docker, Kubernetes with Helm
• Package: pip installable via PyPI

Roadmap

• Prompt CRUD from web UI
• Dark/Light mode
• Tags for cross-category organization
• Section-level favorites
• Sort, Import/Export
• pip installable package
• Usage analytics
• Browser extension

License

MIT License

---

Happy Prompting! 🚀