PromptL is a versatile, user-friendly language that simplifies defining and managing dynamic prompts for LLMs. Whether you’re a developer or a non-technical user, PromptL offers a human-readable format that doesn’t compromise on power or flexibility.
While LLMs are becoming more powerful and popular by the day, defining prompts for them can be a daunting task. All main LLM providers, despite their differences, have adopted a similar structure for their prompting. It consists of a conversation between the user and assistant, which is defined by a list of messages and a series of configuration options. In response, it will return an assistant message as a reply.This structure looks something like this:
Copy
Ask AI
{ "model": "<your-model>", "temperature": 0.6, "messages": [ { "type": "system", "content": "You are an AI assistant that writes personalized birthday messages." }, { "type": "user", "content": "Write a birthday message for Tom, who loves programming!" } ]}
This structure, while straightforward, presents challenges:
Difficult to Write: Non-technical users find JSON hard to write and understand.
Static and Rigid: Simple structures aren’t ideal for dynamic, user-driven conversations.
Code Overhead: Customizing prompts requires repetitive, often messy code.
PromptL solves these issues by offering:Readable and Maintainable Syntax: Easily write and manage prompts.
Dynamic Flexibility: Add dynamic variables to adapt prompts to different scenarios.
Powerful Logic in a Single File: Incorporate logic to handle complex workflows with ease.Here’s the same prompt using PromptL:
Copy
Ask AI
---model: <your-model>temperature: 0.6---You are an AI assistant that writes personalized birthday messages.<user> Write a birthday message for {{ name }} who loves {{ hobby }}!</user>
In this case, not only the syntax is way more readable and maintainable, but it also allows for dynamic generation of prompts by using variables like {{ name }} and {{ hobby }}!
This is just a small example of what PromptL can do. It’s a powerful tool to help you define smarter, more dynamic prompts for your LLMs. Ready to learn more? Let’s dive in!