Write better prompts for ChatGPT, Claude, Gemini, Perplexity, Mistral, Midjourney, and whatever the future may hold.
Structured Prompt is a browser-based tool that makes it easy to create and archive complex structured prompts in a variety of formats. Structured Prompts are platform-agnostic by design, and may be used with any generative AI platform, present and future.
It’s time to get serious about prompting. Structured Prompt will help you write better prompts that produce better responses.
The browser-based Structured Prompt tool will help you engineer prompts that produce responses that are more specific and that are more directly tailored to your intentions. It will also help reduce the errors and misunderstandings that can occur when presenting large prompts in a free-form format. You’ll save an incredible amount of time and will help large language models like ChatGPT, Claude and Gemini return better responses that are aligned with your original intent.
With Structured Prompt Notation you can author once and return your prompt in a variety of formats including JSON and XML
Structured Prompt Notation (SPN) is a simple outline-based format that allows you to easily create complex structured prompts that are automatically converted to JSON, XML, Brackets, and more. No coding is necessary. This lets you author once, and then pick the prompt structure that works best with your large language model of choice.
Moving forward, as new LLMs prefer different structures, our plan is to add new conversion options so that you will not have to rebuild your prompts. SPN will not only help you easily build complex prompts today, it’s designed to help future-proof your prompts for tomorrow.
Sign up as a beta tester, or try Structured Prompt Notation right now using our free SPN Editor.
Use AI-Assist to build prompts using one of our 7 ready-to-use prompt models, or with a custom model that YOU design.
The Structured Prompt AI-Assist feature includes seven ready-to-use prompt models: TRACI, TRACI Interactive, TRI, Expert Guidance, RAFT, Process Loop, and Freeform Prompt Improver. Select a model, describe what you would like to accomplish, and AI-Assist will create a structured prompt based on your input. At that point you can use the prompt as-is, or edit the generated Structured Prompt Notation to make it even better. All structured prompts can be automatically converted from SPN to a variety of formats including JSON, XML, Brackets, and more.
In addition to the seven off-the-shelf models, Structured Prompt includes a Prompt Model Builder that you can use to design your own custom model and then use AI-Assist to generate prompts based on your model.
Structured Prompt Elements are prompt parts that you create once and use across multiple prompts.
Elements are saved prompt parts that can be used again and again in your prompts. Our AI-Assist builder for Elements can be used to create the following parameters: Task, Role, Audience, Create, Intent, Steps, and Variables & Settings. Simply choose a parameter type, describe what your element should include, and AI-Assist will create a reusable parameter that can be added to any prompt.
You can copy and paste the entire text of any element into the Structured Prompt Notation editor. Better still, you can add it as a “Linked Element” so that when you edit the element, all of the prompts that use that element will automatically update to reflect your changes.
In addition to the off-the-shelf element parameters, you can also build your own custom Element Model and use AI-Assist to generate elements that are based on your model.
Ready to try it? Sign Up for a {Structured} Prompt Beta Tester Account Today.
Our platform is phase 2 of beta testing. If you are interested in early access, please sign up to become a Beta Tester. If you participated as a Phase 1 beta tester, please sign up for Phase 2 using the link below.
Meet the TRACI prompt framework for ChatGPT
Task – Role – Audience – Create – Intent
TRACI is a versatile prompting framework that serves as the starting point for the prompts you will engineer in the {Structured} Prompt helper tool for ChatGPT. Individual TRACI elements can be toggled off and on as needed, and each item can include a list of rules and parameters to define and refine the ChatGPT response.
{Task}
Task refers to general type of output that the prompt should achieve (Example: Write marketing copy). Additional Parameters can be added to further define the Task and improve the Response.
{Role}
Role refers to the title or persona of the ‘person’ who will present the response (Example: Professional copywriter). Additional Parameters can be added to further define the Role and improve the Response.
{Audience}
Audience refers to the title or persona of the ‘person’ who the response is designed to reach (Example: Potential Customers). Additional Parameters can be added to further define the Audience and improve the Response.
{Create}
Create provides clear instructions about desired format of the Response (Example: Write 200 words). Additional Parameters can be added to further define the Create statement and improve the Response.
{Intent}
Intent indicates the overall intention or purpose of the Prompt (Example: Persuade the audience to buy a product.) Additional Parameters can be added to further define the Intent and improve the Response.
TRACI is the original framework developed for Structured Prompt and is available in the AI-Assisted prompt builder along with six other powerful prompt models. You can use the Structured Prompt application to build TRACI prompts using AI-Assist, or download our free white paper for a step-by-step guide to building TRACI prompts form scratch.
All {Structured} Prompts are output in both JSON and YAML formats that can be easily understood by ChatGPT. The helper tool generates the JSON prompts for you, so no coding is required.
Structured Prompt Video Tutorials
Lesson 1 – Quickly build a Freeform Prompt and a Structured Prompt
Lesson 2 – Quick overview of Structured Prompt Notation
Lesson 3 – Overview of AI-Assist Prompt Builder Options
Lesson 4 – Quick Look at Prompt Documentation Features
Lesson 5 – Introduction to Prompt Elements (Reusable Prompt Parts)
Lesson 6 – Create an AI Assist Prompt Model Based on an Existing Prompt
{Structured} Prompt Demos
New Feature: With AI-Assist Models, you can now build your own Structured Prompt formats!
AI-Assist for TRACI prompts has long been one of our favorite Structured Prompt features. Now, you can build your own Structured Prompt model and use AI-Assist to create prompts using your model.
How to use Structured Prompt to a create Mini GPT
The Structured Prompt AI-Assist builder can be used to quickly create a powerful expert guidance Mini GPT. In this video we create a Python Coach GPT that will teach users of any experience level to program in Python.
{Structured} Prompt Blog
Introducing TRAACI, our newest AI-assisted prompt model. (The extra A stands for Analyze.)
New Today - Take a quick look at the newest model just added to StructuredPrompt: the TRAACI model. This model bears a close resemblance to the standard TRACI prompt but stands out due to one key feature: it includes an 'Analyze' parameter that instructs the model to...
Announcing AI-Assisted Models for Structured Prompt Elements
Today we are excited to announce a new model-based AI-Assist tool for Elements. As a reminder, Elements are "prompt parts" that can be used again and again across many prompts. Our new feature enables you to use the power of AI to build elements based on predefined...
Introducing the new drag-and-drop Visual Builder for Structured Prompt
This is a quick look at the new drag-and-drop Visual Prompt Builder for Structured Prompt. You can try it now on the Structured Prompt beta site at app.structuredprompt.com. This is a work in progress and currently functions as a standalone prompt builder/editor, but...