PromptRepo

Take back control of your System Prompts. Store them in your repository, not in scattered services.

Note: This project is currently in beta. We welcome feedback and contributions!

The Problem

Managing AI system prompts across environments is painful:

• Prompts scattered across different LLM provider dashboards
• No version control or change history
• Difficult collaboration between developers and product owners
• Infrastructure overhead for each environment
• No easy way to test prompt changes before deployment

The Solution

PromptRepo brings your system prompts into your Git repository where they belong. Treat prompts like code - version them, review them, and deploy them with confidence.

Beta coming soon at beta.promptrepo.dev

Key Features

Chat Playground

Interactive testing environment where anyone - developers, product owners, QA - can chat with your AI agent using the system prompt. See exactly how your agent behaves in real scenarios without writing code.

Prompt Optimizer

AI-powered prompt enhancement that follows provider-specific best practices:

• OpenAI: GPT-4.1+ optimizations, precision patterns
• Anthropic: XML tags, step-by-step thinking
• Google Gemini: Direct, efficient prompting
• OWASP 2025 Guardrails: Optional prompt injection protection

Tool Mocking

Define tools/functions and mock their responses to test how your agent handles different scenarios. Perfect for testing edge cases without calling real APIs.

Evaluations

Run unit tests, semantic similarity checks, and complex evaluations against your prompts:

• Built-in metrics (correctness, hallucination, bias detection)
• Custom evaluation definitions
• Batch execution with history tracking

Share Conversations

Generate shareable links for chat sessions. Show colleagues exactly how your agent performs in specific scenarios - great for demos, reviews, and debugging.

Git-Native Workflow

Every change you make automatically updates your Git branch. Create a PR, iterate on your prompts, and merge when ready. You never lose your work and never worry about syncing - it's just Git.

Multi-Provider Support

Connect your own API keys and switch between providers:

• OpenAI (GPT-4, GPT-4o, etc.)
• Anthropic (Claude family)
• Google (Gemini models)
• Mistral
• And more...

Getting Started

Prerequisites

• Node.js 18+ (for frontend)
• Python 3.11+ (for backend)
• uv - Python package manager (install)
• Git

Quick Start

1. Clone the repository

   git clone https://github.com/YOUR_USERNAME/promptrepo.git
   cd promptrepo

2. Set up environment variables

   cp .env.example .env
   # Edit .env with your configuration

3. Start the backend

   cd backend
   uv sync
   uv run uvicorn main:app --reload --port 8080

4. Start the frontend (in a new terminal)

   cd frontend
   npm install
   npm run dev

5. Open your browser Navigate to http://localhost:3000

Docker Development

docker-compose -f docker-compose.dev.yml up

Project Structure

promptrepo/
├── backend/                 # Python FastAPI backend
│   ├── api/v0/             # Versioned API endpoints
│   ├── services/           # Business logic layer
│   ├── agents/             # AI agent implementations
│   ├── database/           # SQLModel ORM & migrations
│   └── tests/              # Unit & integration tests
├── frontend/               # Next.js React application
│   ├── app/                # Pages (App Router)
│   ├── stores/             # Zustand state management
│   ├── services/           # API client services
│   └── components/         # Reusable UI components
├── docs/                   # Architecture documentation
└── persistence/            # Shared storage for Git repos

Tech Stack

Backend

• FastAPI - Modern async Python web framework
• SQLModel - ORM combining SQLAlchemy + Pydantic
• GitPython - Git repository operations
• DeepEval - LLM evaluation framework
• any-llm / any-agent - Multi-provider LLM abstractions

Frontend

• Next.js 15 - React framework with SSR
• React 19 - UI library
• Zustand - Lightweight state management
• Chakra UI - Component library
• TypeScript - Type safety

Configuration

Environment Variables

Variable	Description
DATABASE_URL	Database connection string
GITHUB_CLIENT_ID	GitHub OAuth app client ID
GITHUB_CLIENT_SECRET	GitHub OAuth app secret
HOSTING_TYPE	INDIVIDUAL or ORGANIZATION
REPO_PATH	Path to prompt repository storage

See .env.example for all available options.

Running Tests

Backend

cd backend
uv run pytest

Frontend

cd frontend
npm test

Roadmap

We're actively working on:

• Auto-migration tooling - Generate YAML metadata from existing repos for quick PromptRepo adoption
• CI/CD integration - Run prompt evaluations in your pipeline
• Prompt versioning UI - Visual diff and history
• More evaluation metrics - Custom metric definitions
• Prompt templates - Reusable prompt patterns

Contributing

We love contributions! Please see CONTRIBUTING.md for guidelines.

Before submitting a PR:

1. Ensure tests pass (uv run pytest for backend, npm test for frontend)
2. Follow existing code patterns
3. Update documentation if needed

Feedback

This is a beta release - your feedback shapes the future of PromptRepo!

• Try it out: beta.promptrepo.dev (coming soon)
• Report issues: GitHub Issues
• Share feedback: We'd love to hear how you're using PromptRepo

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Stop managing prompts in scattered dashboards. Start managing them like code.