# Prompt Sapper: A LLM-Empowered Production Tool for Building AI Chains

**Authors:** Yu Cheng, Jieshan Chen, Qing Huang, Zhenchang Xing, Xiwei Xu, Qinghua Lu  
**arXiv:** [2306.12028](https://arxiv.org/abs/2306.12028)  
**Date:** 2023-06-21  
**Summary page:** https://www.summarizepaper.com/en/paper/2306.12028v1/

## Key points

- Introduction of AI chains and methodology for AI chain engineering
- Existing methods like chatbots are not suitable for reusable AI services
- APIs like LangChain require significant programming knowledge, creating a barrier
- Proposal of Prompt Sapper as a no-code integrated development environment for building AI chains
- Use of chat-based requirement analysis and visual programming in Prompt Sapper
- User study demonstrating improved performance and quality of AI chains with Prompt Sapper
- Ability to leverage foundation models effectively without extensive programming knowledge
- Innovative approach to utilizing foundation models for building reusable AI services
- Opening up new possibilities in various domains by enabling efficient communication with advanced language models.

## Summary

The paper titled "Prompt Sapper: A LLM-Empowered Production Tool for Building AI Chains" introduces the concept of AI chains and proposes a methodology to systematize AI chain engineering. The emergence of foundation models, such as GPT-4 and DALL-E, has revolutionized various domains by enabling communication with AI through natural language prompts. However, existing methods like chatbots are not suitable for building reusable AI services. While APIs like LangChain allow for LLM-based application development, they require significant programming knowledge, creating a barrier for many users. To address this issue, the authors propose the concept of AI chains and incorporate principles and practices from software engineering into AI chain engineering. They develop Prompt Sapper, a no-code integrated development environment that embodies these engineering principles and patterns naturally during the process of building AI chains. Prompt Sapper allows AI chain engineers to compose prompt-based AI services on top of foundation models using chat-based requirement analysis and visual programming. The paper highlights a user study that evaluated Prompt Sapper's efficiency and correctness. The results demonstrate that Prompt Sapper improves the performance and quality of AI chains. With this tool, users can leverage foundation models effectively without requiring extensive programming knowledge. Overall, "Prompt Sapper: A LLM-Empowered Production Tool for Building AI Chains" presents an innovative approach to overcome the barriers in utilizing foundation models for building reusable AI services. By introducing the concept of AI chains and developing an intuitive development environment like Prompt Sapper, this research opens up new possibilities in various domains by enabling efficient communication with advanced language models.