PromptWorks is a full-stack solution designed for managing and testing prompts used with large language models. It targets teams involved in LLM operations and prompt engineering, offering a unified platform for prompt lifecycle management, version control, and collaborative experimentation, thereby streamlining prompt development and deployment workflows.

How It Works

The platform employs a robust backend built with Python, FastAPI, and SQLAlchemy, leveraging PostgreSQL for data persistence and Redis for caching and task queuing via Celery. The frontend is developed using Vite, Vue 3, and Element Plus, providing an interactive user interface. This architecture supports prompt creation, versioning, categorization, and detailed comparison views. It also centralizes model configuration and quota management, enabling A/B testing and evaluation experiments through backend execution and a dedicated frontend testing panel.

Quick Start & Requirements

Deployment is recommended via Docker Compose for a streamlined setup, pulling pre-built backend-main-latest and frontend-main-latest images. Alternatively, local development requires Python 3.10+, Node.js 18+, PostgreSQL, and Redis. Key setup steps involve synchronizing backend dependencies (uv sync --extra dev), initializing environment variables (cp .env.example .env), setting up the database, installing frontend dependencies (cd frontend && npm install), and running backend (uv run poe server) and frontend (npm run dev -- --host) services. The backend API is accessible at http://localhost:8000/api/v1 (with docs at /docs), and the frontend at http://localhost:18080 or http://localhost:5173 for local dev. Multi-architecture Docker images are available for Apple Silicon/ARM devices.

Highlighted Details

• Comprehensive Prompt Lifecycle Management: Supports creation, version iteration, and tag-based categorization with audit trails.
• Version Comparison: Features a diff view to quickly identify changes between prompt versions.
• Model Operations: Centralized management of LLM services and call quotas for A/B testing.
• Evaluation & Testing: Backend capabilities for experiment execution and metric recording, integrated with a frontend testing panel.
• Multi-LLM Support: Configuration for integrating services like OpenAI and Anthropic via API keys.

Maintenance & Community

The project outlines a clear contribution guide emphasizing a "format -> lint -> test" workflow using uv run poe test-all. It welcomes Issues and suggestions for improvement, fostering a collaborative development environment. Specific community channels (e.g., Discord, Slack) or core maintainer details are not explicitly provided in the README.

Licensing & Compatibility

The provided README does not specify a software license. This absence represents a significant adoption blocker, as the terms for use, modification, and distribution remain undefined, potentially restricting commercial or closed-source integration.

Limitations & Caveats

The current frontend example utilizes mock data, indicating that full API integration may require further development or configuration. The lack of explicit licensing information is a critical caveat for any potential adopters.