claude-prompt-generator  by aws-samples

Prompt generator for Claude models

Created 1 year ago
1,271 stars

Top 31.2% on SourcePulse

GitHubView on GitHub
1 Expert Loves This Project
Project Summary

This project provides a tool for generating and refining prompts specifically for Anthropic's Claude 3 language model, assisting users who are new to Claude or migrating from other models like GPT. It aims to improve prompt effectiveness through an iterative evaluation and revision process, ultimately enabling users to achieve desired outputs with Claude 3.

How It Works

The tool facilitates two primary workflows: generating a Claude 3 prompt from scratch based on Claude 3's best practices (e.g., XML tags) or translating an existing GPT prompt into a Claude 3-compatible format. It then incorporates an auto-evaluation and manual refinement loop, allowing users to compare Claude 3 outputs against original prompts, provide feedback, and iterate on prompt revisions until satisfactory results are achieved.

Quick Start & Requirements

  • Installation: pip install -r requirements.txt
  • Prerequisites: AWS account with AWS CLI and configured credentials for AWS Bedrock API access. An OpenAI API key is optional for prompt evaluation features.
  • Configuration: Copy .env.example to .env and fill in OPENAI_API_KEY, OPENAI_API_URL (optional), and REGION_NAME (e.g., us-east-1).
  • Run: Navigate to the src directory and run python app.py. Access the UI at the provided URL (e.g., http://127.0.0.1:7860).
  • Documentation: AWS CLI: https://aws.amazon.com/cli/, OpenAI API: https://platform.openai.com/docs/api-reference

Highlighted Details

  • Supports prompt generation from scratch or translation from OpenAI prompts.
  • Features an iterative prompt evaluation loop with manual and auto-evaluation options.
  • Includes an experimental feature for generating SEO-optimized product descriptions.
  • Utilizes AWS Bedrock for Claude 3 API access.

Maintenance & Community

No specific details on contributors, sponsorships, or community channels are provided in the README.

Licensing & Compatibility

  • License: Apache-2.0 License.
  • Compatibility: Permissive license suitable for commercial use and integration with closed-source projects.

Limitations & Caveats

AWS Bedrock API access is currently limited to specific AWS regions, which may restrict usability depending on the user's location. The prompt evaluation process involves a manual "10% human-in-the-loop" component for production-ready quality.

Health Check
Last Commit

11 months ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
0
Star History
17 stars in the last 30 days

Explore Similar Projects

Starred by Clement Delangue Clement Delangue(Cofounder of Hugging Face), Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems"), and
10 more.

promptsource by bigscience-workshop

0.2%
3k
Toolkit for creating, sharing, and using natural language prompts
Created 4 years ago
Updated 1 year ago
Feedback? Help us improve.