Self-Hosting Guide

Capabilities

Curated Prompt Library — Access 100+ high-quality, community-tested prompts for ChatGPT, Claude, Gemini, Llama, Mistral, and other AI models
Discover & Browse — Explore prompts by categories, tags, or AI-powered semantic search
Create & Share Prompts — Submit your own prompts with support for text, structured (JSON/YAML), and media-enhanced formats
Version Control — Track prompt changes with built-in versioning and change request system (similar to PRs)
Personalized Feed — Subscribe to categories and get a curated feed of prompts matching your interests
Private Prompts — Keep your prompts private or share them with the community
Voting & Leaderboard — Upvote prompts and discover the most popular ones via PromptMasters leaderboard
Multi-language Support — Available in English, Spanish, Japanese, Turkish, and Chinese

Benefits

Unlock AI Potential: Stop struggling with prompt engineering — use battle-tested prompts from 141k+ GitHub stars community
Save Time: Copy prompts with one click, customize variables inline, and use them instantly in any AI chat
Community-Driven Quality: Every prompt is curated and refined by the community through change requests and voting
Self-Hostable: Deploy your own white-labeled prompt library for your team or organization with customizable branding, themes, and authentication
CC0 Licensed: All prompts are public domain — use them freely for any purpose, commercial or personal

Getting Started

Requirements:

Plan: Free and open-source (CC0 license)
User Permissions: No account needed to browse; sign in via GitHub/Google to create & save prompts
Availability: Generally Available at prompts.chat

This guide explains how to deploy prompts.chat on your own private server for enhanced privacy and customization.

Prerequisites

Node.js 18+
PostgreSQL database
npm or yarn

Environment Variables

Create a .env file based on .env.example:

# Database
DATABASE_URL="postgresql://user:password@localhost:5432/prompts"

# Authentication (choose one provider)
# GitHub OAuth
AUTH_GITHUB_ID="your-github-client-id"
AUTH_GITHUB_SECRET="your-github-client-secret"

# Or Google OAuth
AUTH_GOOGLE_ID="your-google-client-id"
AUTH_GOOGLE_SECRET="your-google-client-secret"

# NextAuth
AUTH_SECRET="generate-a-random-secret"

# Optional: AI-powered semantic search
OPENAI_API_KEY="your-openai-api-key"

Installation

Quick Start (Recommended)

The fastest way to create a new prompts.chat instance:

npx prompts.chat new my-prompt-library
cd my-prompt-library

This will:

Clone a clean copy of the repository (without development files)
Install dependencies
Launch the interactive setup wizard

Manual Installation

Clone the repository

git clone https://github.com/f/prompts.chat.git
cd prompts.chat

Install dependencies
```
npm install
```
Run the interactive setup wizard
```
npm run setup
```
This will guide you through configuring:
- Branding — App name, logo, description
- Theme — Primary color, border radius, UI variant
- Authentication — GitHub, Google, Apple, Azure AD, or email/password
- Languages — Select from 16 supported locales
- Features — Private prompts, categories, tags, comments, AI search, AI generation, MCP support

Configure environment variables

cp .env.example .env
# Edit .env with your database and auth credentials

Run database migrations
```
npm run db:migrate
```
Seed initial data (optional)
```
npm run db:seed
```
Start the development server
```
npm run dev
```
Build for production
```
npm run build
npm run start
```

Configuration

The setup wizard (npm run setup) generates prompts.config.ts automatically. You can also manually edit it:

// Set to true to use your own branding instead of prompts.chat branding
const useCloneBranding = true;

export default defineConfig({
  // Branding
  branding: {
    name: "Your Prompt Library",
    logo: "/your-logo.svg",
    logoDark: "/your-logo-dark.svg",
    description: "Your custom description",
  },

  // Theme
  theme: {
    radius: "sm",        // "none" | "sm" | "md" | "lg"
    variant: "default",  // "flat" | "default" | "brutal"
    colors: {
      primary: "#6366f1",
    },
  },

  // Authentication
  auth: {
    provider: "github",  // "credentials" | "github" | "google" | "azure"
    allowRegistration: true,
  },

  // Features
  features: {
    privatePrompts: true,
    changeRequests: true,
    categories: true,
    tags: true,
    aiSearch: false,  // Requires OPENAI_API_KEY
  },

  // Homepage
  homepage: {
    useCloneBranding,  // Use your branding on homepage
    achievements: {
      enabled: !useCloneBranding,  // Hide prompts.chat achievements
    },
    sponsors: {
      enabled: !useCloneBranding,  // Hide prompts.chat sponsors
    },
  },

  // Internationalization
  i18n: {
    locales: ["en", "es", "ja", "tr", "zh"],
    defaultLocale: "en",
  },
});

Clone Branding Mode

When useCloneBranding is set to true, the homepage will:

Display your branding name as the hero title
Show your branding description below the title
Use your logo as a watermark background instead of the video
Hide the "Deploy Your Private Server" button
Hide the achievements section (Forbes, GitHub stars, etc.)
Hide the sponsor links and "Become a Sponsor" CTA

This is ideal for organizations that want to deploy their own white-labeled prompt library without prompts.chat branding.

Support

For issues and questions, please open a GitHub Issue.

5.9 KiB Raw Blame History