Markdown Prompting In AI Prompt Engineering Explained - Examples + Tips

Prompting is not at all dead in 2025. To create custom GPTs or AI agents – you still need to write effective prompts.

One prompt engineering hack I knew yet never took seriously was the use of markdown prompting for AI models.

Why? – I was too lazy to convert my prompt into markdown text.

Recently, I tried making custom GPTs on the ChatGPT platform. I noticed a significant difference between the markdown prompt and the normal text prompt. I noticed the prompts written on AI Agents platforms were in markdown format.

Hence, I decided to dig deeper about markdowns and understand how to use them in better prompt engineering. Here’s what you will learn:

What is markdown prompting for AI prompt engineering
Markdown elements and how they help improve AI prompt writing

Tips on using markdown prompting

What is markdown text?

Markdown is a simple way to add style to plain text. It uses special characters like ‘#’ for big titles and ‘-‘ for lists. This makes your text look neat and clear.

Basically, you write normally. But then you add a few marks to show that one line is a title or that a list is starting.

For example, the simplest markdown is a ‘list‘.

Use a dash – or an asterisk * to start a list item. This tells the reader that you are listing things.

Example:

- Apples starts a bullet list.

There are many other syntax for markdown, which I will explain in coming sections of this guide.

Markdown text is very simple. You do not need fancy programs to use it. You write in a text editor and then the computer changes it into pretty web pages or documents. This means that markdown is both easy for people to write and easy for computers to understand.

Pinterest: Pros and Cons of using Markdown text for AI prompt engineering

Many websites, like GitHub and blogs, use markdown because it is fast and simple. When you use markdown, you save time, and your writing stays clean and clear.

Difference between markdown, plain text, and HTML for AI prompt writing

If you are still confused, here’s a detailed comparison to help you visualize how each text format gives different output. I have also shared why each format is not practical for AI prompt engineering:

Plain vs Markdown text format

Plain text has no extra symbols. You just write your ideas as you normally speak. The problem arises when you list many ideas in plain text. The AI may have a hard time knowing where one idea ends and another begins.

Here’s an example of how plain text looks like:

HTML vs Markdown text format

HTML uses tags like <h1>, <p>, and <li> to mark sections, paragraphs, and lists. It is very detailed and gives a very strict structure.

While this structure is clear, HTML often makes the text longer. It adds extra parts that the AI might not need. This can slow down the processing because the AI must read through all the extra tags.

Here’s how HTML is different from markdown:

How HTML is different from markdown text

As you can see, HTML wraps text in opening and closing tags. The snippet shows an <h1> tag wrapping “Heading 1” and a <p> tag wrapping the paragraph text. Both opening and closing tags are required to define the structure.

The same heading is written using the # symbol, and the paragraph is simply written as plain text. There is no need for closing symbols, which makes Markdown less verbose and easier to write.

Why use markdown text for AI models?

The more you make it easy for AI to understand your prompt, the better answer it gives. Markdown text is a means to achieve this simplicity with below key benefits:

Easy to Read: Markdown shows your ideas in clear sections. For example, you can use a heading to tell the AI where a new idea starts.
Simple Rules: You only need a few symbols to make your text look organized. This simplicity helps both people and AI understand your message.
Boosts Clarity: When you use markdown in your prompts, you help the AI see the structure of your instructions. This leads to better answers.

Think of AI as a super-smart helper that needs clear instructions to give you the best answers.

Prompt engineering is just a fancy term for creating these clear instructions.

Markdown is like a secret code. It uses simple symbols, like # for big titles or * for making words bold. This code makes your instructions look neat and easy to follow.

Guide to core markdown elements for AI prompts

Now, I will explain the many markdown text formatting elements.

1. Headings

Headings use the “#” symbol. The more “#” symbols you add, the smaller the heading becomes.

Why us ‘headings’ markdown important for AI models?

A clear heading tells the AI, “Look here for important info!”

For example:

H1 (Heading 1): Use one “#” for the biggest title. This is like the title of a book.
H2 (Heading 2): Use two “##” for main sections. Think of these as chapter titles.

H3 (Heading 3): Use three “###” for subsections. They are like section titles inside a chapter.

Diagram showing how different markdown heading elements look like

2. Bold and Italics

Bold and italics help your text stand out. Bold text looks thicker and grabs attention, like a flashing sign. Italics make words lean to the side, like a gentle whisper that draws your eye.

Why ‘bold and italics’ are important for AI models?

Use bold when you want to shout, “This part matters!” and italics when you want to hint, “This part is special” without overwhelming the rest of the text.

Here’s a table showcasing the format and how AI models perceive them:

Markdown Element	Plain Text	Bold Text	Italics Text
Example	This is plain text	This is bold text	This is italics text
Format	NA	Use two asterisks like `bold`	Use one asterisk like `italic`
How AI models perceive the element	NA	It tells the AI that these words are very important.	It shows that the word is special or a little extra.

3. Lists

Lists help you break down information into clear, simple steps. They work like a grocery list. It tells you what to buy. They also work like a checklist, showing you what to do next.

There are two main types of lists: unordered lists and ordered lists.

Unordered Lists

Unordered lists use symbols like dashes ( – ) or asterisks ( * ) to mark each item. They don’t need represent a special order. For example, here is a list of fruits, with no order of preference:

Apple
Banana
Cherry

This type is great when the order of the items does not matter. Think of it as a simple list of your favorite fruits.

Ordered Lists

Ordered lists use numbers followed by a period. They show a step-by-step order.

Example:
1. Brush your teeth.
2. Get dressed.
3. Eat breakfast.

Ordered lists work like a recipe or a set of instructions. The order is important because each step comes after the other.

Nested Lists

Sometimes, you need to make lists inside lists. This is called nesting. It is useful when you want to show a main idea and then break it down into smaller parts.

Example:
1. Make a sandwich.
  - Get bread.
  - Add cheese.
  - Put on lettuce.
2. Enjoy your meal.

Nested lists help you show details under one main point.

Why is ‘lists’ markdown important for AI models?

Lists organize ideas so the reader or the AI can follow them easily. They make your prompt clear and your instructions simple. This helps the AI understand what to do and makes your answers better.

How to design ‘lists’ as markdown text for AI models?

This might seem basic – but there is a framework that I think you must know for creating a list. The more you know, better!

Start: Begin by thinking about your main idea.
Decide on List Type: Ask yourself if the order of items is important.
- If yes, choose an ordered list.
- If no, choose an unordered list.
Write Items: For an ordered list, write the items with numbers. For an unordered list, use dashes or asterisks.

Check Clarity: Review your list to make sure every item is clear.
Nest Sub-Items: If you need more details, add nested items under a main item.
Final List: Your organized list is ready!

Here’s a list I designed to help you visualize how the above instruction must be represented to AI models like ChatGPT:

flowchart to make a list:

    A[Start: Think of your main idea] --> B[Decide on the list type]
    B --> C{Is order important?}
    C -- Yes --> D[Choose Ordered List]
    C -- No --> E[Choose Unordered List]
    D --> F[Write items in order (1, 2, 3...)]
    E --> F[Write items using symbols (- or *)]
    F --> G[Check each item is clear]
    G --> H{Need more details?}
    H -- Yes --> I[Nest sub-items under main item]
    H -- No --> J[Final List Ready]

Using lists in your AI prompts makes your instructions neat and easy to follow. This helps the AI give you better and more accurate answers.

4. Code Blocks

Code blocks keep special text, like computer code, neat and in one place.

It let you show text exactly as you want it to. It does not allow the computer change any of the spaces or symbols. Code blocks are like a framed picture; they protect the text from getting mixed with other words.

When you use code blocks, your text is shown in a “monospaced” font—just like what you see in computer programs. This makes it easier to read code or special commands.

Inline Code:

For this, use a single backtick (`) around a short snippet of text.

Here is some code

When you do this, the AI sees this text as code. It shows it in a different font as done above.

Multiline Code Blocks:

Use three backticks (“`) before and after your text.

For example, imagine you have a block of text that you want to show as code. Here’s how it looks without and with code formatting:

Before Code Formatting:
Here is a block of text that shows some code:
def greet():
print(“Hello, world!”)
greet()

When formatted as a code block, it keeps the spacing and symbols exactly as you wrote them.

After Code Formatting (Using a Code Block):

def greet():
    print("Hello, world!")
greet()

In the “before” version, the text might look messy and the spacing could be lost or changed when viewed. In the “after” version, the text stays exactly the same. It maintains the correct spaces and line breaks. This consistency makes it much easier to read.

5. Links

Links connect your text to another web page. It is like a magic doorway to take you from one place to another on the web.

By linking text, you make it easy for both people and AI to jump to another page for more information.

Why use links markdown for AI models?

Guides the Reader: Links help readers find extra information.
Organizes Information: They make your text neat, like putting a label on a box that tells you what’s inside.
Saves Time: You don’t need to write a lot of extra details. Instead, you can just send people to the full story on another page.

These same benefits apply for AI models too.

By using links in your AI prompts or any text, you help guide the reader. You show the AI exactly where to go. This makes your message clear and useful.

Good links make your prompt like a treasure map, guiding the AI to extra information.

How to create a link:

First, place Text in Brackets. Write the words you want to show as a clickable button inside square brackets.

Example: [Click Here]

Then, add URL in Parentheses. Right after the brackets, write the web address (URL) inside round parentheses.

Example: (https://www.appliedai.tools)

When you combine them, you get:
[Click Here](https://www.example.com)

This tells the AI and your readers, “When you click this, go to the website!”

Below is a simple diagram to show how a link works. Imagine the text is inside a box that leads to a website:

  +-----------------------+
  |     [Click Here]      |  --->  https://www.appliedai.tools
  +-----------------------+

The box contains the text [Click Here]. When you click the box, it opens the URL https://www.appliedai.tools.

Use clear words in the brackets – try using words that tell your reader what to expect. For example, [Learn More About Space](https://www.isro.gov.in/) tells your reader that they will learn about space if they click the link.

6. Images

Images in Markdown let you add pictures to your text. They help make your prompt or webpage look fun and clear.

You use an exclamation mark followed by square brackets for a short description. This is called “alt text.”

Then, you add parentheses with the image URL. You can also add a title, which shows up as a tooltip when you hover your mouse.

Why use Images Markdown for AI models?

Images make your prompt visual and can help the AI understand your ideas better. It can turn abstract ideas into clear pictures.

AI models like ChatGPT can open links. They are also multi-modal – meaning, you can separately add images to enrich your prompt.

How to Insert an Image:

Basic Image Tag Format:
![Alt text](image-url "Optional Title")

Use an exclamation mark, then brackets for alternative text, and parentheses for the image URL.

Alt text: A short description of the image. This is important because it tells people what the image is about. It helps if the picture does not load or for disabled users.
Image URL: The web address where the image is located.

Optional Title: Extra text that appears when you move your mouse over the image.

For example, if you have an image of a cute cat, here’s how you can write it in markdown format:

![Cute Cat](https://example.com/cat.jpg)

7. Line breaks and horizontal rules

Line breaks help separate ideas, while horizontal rules (three dashes) make a clean break between sections.

Line Breaks

Line breaks tell the computer to move text to a new line. You add a line break by putting two spaces at the end of a line and then pressing Enter.

Before a line break:
“This is the first sentence. This is the second sentence.”
(They show as one long line.)

After adding a line break:
“This is the first sentence.
This is the second sentence.”
(Now, the sentences appear on separate lines.)

Using line breaks helps the text look neat. It is similar to writing a poem where each line is on a new row.

Horizontal Rules

A horizontal rule is a solid line that separates sections of text. You make one by typing three dashes (---) on a new line.

Before a horizontal rule:
“Section One: This part talks about animals.”
“Section Two: This part talks about plants.”
(The text runs together.)

After adding a horizontal rule:
“Section One: This part talks about animals.” “Section Two: This part talks about plants.”
(The rule shows as a line between the two sections.)

A horizontal rule works like a divider in your notebook. It tells the reader, “This is the end of one topic, and a new topic begins here.”

Here’s how it looks visually:

Before adding a Line Break and Horizontal Rule, the text appears as one long block:

Image showcasing before adding a Line Break and Horizontal Rule

After adding a Line Break and Horizontal Rule, the sentences split into separate lines. A clear line divides the sections:

Image shows output after adding a Line Break and Horizontal Rule

Why add lines or horizontal rules for AI prompts?

A line break is a point where the text moves to a new line. A horizontal rule is a line that divides content.

When you add a line break or horizontal rule, you help the AI and your readers see your text clearly. It makes instructions and information easy to follow, like clear steps in a recipe.

8. Tables

Tables in Markdown help you organize data into rows and columns, just like a simple grid or a small spreadsheet. They let you compare items and keep information neat.

Why use tables for AI prompts?

Tables are like a game board with boxes that hold your data. They make it easier for the AI to understand and compare different pieces of information.

Use tables for clear, organized data. This includes a list of items, scores, or other details that fit into columns and rows. Tables give your prompts extra structure, making your instructions clearer for the AI and helping you get better, organized responses.

How to create a table:

To make a table, use the pipe symbol (|) to separate columns. Use dashes (–) to mark the header row. For example, type this:

| Name  | Age |
|-------|-----|
| Alice | 10  |
| Bob   | 12  |

When rendered, it looks like this:

Name	Age
Alice	10
Bob	12

How markdown prompting improves AI model responses?

We understood why one should use markdown prompting for AI models – for better responses.

I also explained various markdown elements and how each help with AI prompt optimization.

Now, I will help you understand how does markdown help structure the prompts when the elements are put together.

Consider the below diagram – check the ‘messy prompt’. This is how one would usually type their prompts to AI models like ChatGPT, Claude, or DeepSeek.

Now check the ‘structured markdown prompt’. It uses the ‘headings’ and ‘lists’ – and that itself changes the prompt so much.

Here are key takeaways from the above prompt comparison:

1. Clear structure leads to clear results

When you use Markdown, you break your prompt into sections with titles, bullet points, and lists. This structure acts like a map that tells the AI where each idea begins and ends.

2. Reduces confusion with simple rules

Markdown removes guesswork for the AI. When you write a clear header or list, the AI understands exactly what to do.

For example, a header like “## Steps” tells the AI to expect a list of steps.

Flowchart showcasing clear markdown prompting flow which helps structure the AI prompt

3. Saves time with short and sharp prompts

Short prompts with clear Markdown symbols use fewer words. This saves the AI from reading extra text and helps it focus on what matters.

In AI terms, “tokens” are like words. Fewer tokens mean the AI can work faster and with less chance of error.

Plan text versus markdown enhanced prompting comparison - in terms of token counts.

4. Consistent format boosts confidence

When you format your prompt the same way every time, the AI learns your style. This consistency helps it give answers that always follow the same structure.

Can I convert simple text into markdown prompting format?

You can write the prompt as required, and use these below hacks to convert the plain text into markdown prompt format:

Write your prompt with required formatting on Google Docs. Now, download the document and share this as an input for AI models like Claude, ChatGPT, DeepSeek. Then ask it to convert the document text into corresponding markdown text.
Use freely available online text editors Rich Text to Markdown, tobwil on GitHub, etc. Simply Google and you will find many such online tools.

If you want to make markdown your lifestyle, here are some online tools to convert plain text into Markdown formatting:

Tool/Editor	Description	Platform
Dillinger	A simple online Markdown editor that shows a live preview of your formatted text as you type.	Web-based
StackEdit	A robust online Markdown editor that lets you write and save documents, sync with cloud storage, and collaborate.	Web-based
Markable	An easy-to-use online tool for converting plain text into Markdown and vice versa.	Web-based
Typora	A desktop Markdown editor with a seamless live preview and minimal interface, ideal for distraction-free writing.	Windows, Mac, Linux
Visual Studio Code	A powerful code editor that supports Markdown editing with built-in preview features and various extensions for enhanced functionality.	Desktop (Windows, Mac, Linux)
Obsidian	A note-taking app that uses Markdown for formatting, great for creating and managing structured notes and documents.	Desktop & Mobile

Here is a good resource I found to learn more about markdown in detail – Markdown Guide

3 tips on using markdown prompt engineering techniques

I have already covered a few advanced prompt engineering techniques in the past, and would highly recommend learning them:

Learn how to use Tree of Thoughts Prompting with examples.

Learn how to use Chain of Thoughts prompting with examples.
Learn how to write an AI prompt using best practices.

For this section, I will focus on sharing prompting tips for using markdown to write AI prompts:

1. Ask the AI to reply in markdown format

Clearly instruct the AI to respond using Markdown formatting.

For example, say, “Please write your answer in Markdown, with a title, a bullet list, and a conclusion.”

This tells the AI to use the same clear, organized style when it answers.

2. Combine multiple markdown prompting techniques in a single prompt

Write your entire prompt using headers, lists, block quotes, code blocks, and emojis. This makes your prompt look like a well-organized document.

Imagine you want the AI to generate an environmental report. You want it to have a title, an introduction, a data summary, and a conclusion.

Here’s how you could write that prompt using multiple Markdown techniques:

# Environmental Impact Report

> **Note:** Use clear and simple language. Do not add extra details beyond what is asked.

## Introduction
- Write a short paragraph explaining the impact of climate change.
- Include one **bold** example of a recent climate event.

## Data Summary
- Present the following data in a code block as JSON:
{ "location": "Global", "year": 2024, "effects": ["Rising temperatures", "Melting ice caps", "Severe storms"] }

- Then, list three key points from the data:
  - Rising temperatures are affecting ecosystems.
  - Melting ice caps contribute to sea-level rise.
  - Severe storms increase natural disasters.

## Conclusion
- Write one sentence that summarizes the report.

Header (# Environmental Impact Report): Sets the title.
Block Quote (> Note: …): Emphasizes an important instruction.
Subheaders (## Introduction, etc.): Divides the prompt into clear sections.

Bullet Lists: Tells the AI what to include in each section (steps or items).
Code Block for JSON Data: Shows the exact structure for the data summary.
Final Bullet List: Reminds the AI to list specific points.

3. Use footnotes for extra details

Footnotes let you add extra information without crowding your main prompt text. You insert a small marker where extra detail is needed. Then, at the bottom of your prompt, you add the full explanation. This keeps your main instructions neat and clear.

In your Markdown, write a sentence and add a footnote marker like this:

Please list your favorite books.[^1]

Then, at the bottom of your prompt, include the extra details:

[^1]: Include only the books you have read in the last year.

Thus, any extra instructions or clarifications don’t interrupt the flow of your main task.

You can give background info or examples in a non-intrusive way. For example, if a term is complex, a footnote can explain it without slowing down the main text.

Where to use markdown prompting for AI models? – Markdown examples

By now you must have realized that markdown prompting is indeed extra work. Sometimes, you can just prompt casually and get your job done. But some cases do give better results with markdown prompting – here are a few of them:

1. Create content

A lot of professional writers and coders use markdown formatting to organize their own writing. Same goes for AI models – it helps organize your writing.

When you write a blog post, you can use one hash (#) for a big title. Use two hashes (##) for subheadings. Use dashes (–) for bullet points. For example, your prompt might say: “Write a blog post about healthy eating. Use a big title, clear subheadings, and list tips at the end.”

2. Technical documentation and code

Coding is another key use case for AI models, wherein you can generate lines of code without much efforts. I too used Claude to generate a few images on this blog post you are reading!

Markdown formatting helps the AI to format code neatly. Coding is a game of logic. One has to be precise and clear with the expected output and process to get good results.

You can ask it to write or fix code. AI models will use “code blocks” (text surrounded by three backticks) to show the code.

P.S. A code block is a section of text that looks like computer code. It makes it easier to read and copy.

3. Use of chatbots and virtual assistants

When you use Markdown in your prompt, the AI (such as a chatbot) gives answers that are neat and easy to read. This means your conversation feels more like a well-organized list or set of instructions. These chatbots include ChatGPT, Claude, DeepSeek, etc.

You can tell the AI: “List three things you can do at a park. Use bullet points to separate each item.”

A chatbot is a computer program that talks with people. Using Markdown helps the chatbot separate different ideas.

4. Extract data and generate reports

Markdown can make tables and lists. This is useful when you want the AI to pull data from a text. It can show the information in a clear format, like a report.

For example, ask the AI: “Show me a table with three columns: Name, Age, and City. Fill in three sample rows.” AI models can also go through your data and help generate insight.

5. Study planning and scheduling

A schedule or study plan is a list of tasks organized by time. Use Markdown to plan a study schedule. It helps break down the plan into sections.

This helps the AI list out tasks in an order that makes sense.

You can prompt the AI: “Make a study plan for a math test. Use a heading for each day and list three topics under each day.”

6. Meal planning and recipes

A meal plan is a plan that tells you what to eat at different times of the day.

Markdown can help organize recipes and meal plans. Using headers and bullet points makes it clear. You can ask the AI to list meals for breakfast, lunch, and dinner.

Your prompt might say: “Create a meal plan for one day. Use a header for each meal and list two recipes under each meal.”

7. Product reviews and feedback

Pros and cons if often listed for product reviews, which mean the good things and bad things about a product.

Use Markdown to organize product reviews. Markdown helps show these in a list. The AI can help list the good and bad points of a product in a clear, bullet-pointed list.

Ask the AI: “Write a review for a new smartphone. Use bullet points for the pros and cons.”

Frequently Asked Questions (FAQs) about markdown prompting

What is markdown in ChatGPT?

Markdown is a simple way to add formatting to plain text, making it easy to create structured documents. In ChatGPT, Markdown is used to format responses, allowing for elements like headings, lists, code blocks, and tables.

For example, when ChatGPT includes code in its replies, it uses Markdown to show the code in a formatted block. This makes it easier to read. While ChatGPT can generate text using Markdown, the ChatGPT interface is responsible for how this formatted text appears visually.

What format is ChatGPT text?

ChatGPT generates responses in plain text, enriched with Markdown formatting to structure content effectively. This approach allows for the inclusion of various elements. These include headings, lists, code blocks, and tables. This enhances the clarity and organization of the information presented.

What does “` mean in markdown?

In markdown prompting, enclosing text with triple backticks (“`) creates a code block. This is a section where code or text is displayed in a monospaced font. It preserves formatting and indentation.

Can ChatGPT underline text?

In ChatGPT’s standard interface, underlining text is not supported through Markdown formatting. ChatGPT can generate HTML-formatted text. It can also generate markdown formatted text. The rendering of such formatting depends on the platform or application displaying the content.

Who created Markdown?

Markdown was developed by John Gruber in collaboration with Aaron Swartz. Introduced in 2004, it is a lightweight markup language. It is designed to be easy to read and write. This allows users to format plain text using simple syntax. The syntax can be converted into HTML.

Is markdown better than HTML?

The choice between Markdown and HTML depends on the specific requirements of your project. Markdown is suitable for simpler formatting needs, while HTML is necessary for more complex web development tasks.

Here’s a concise comparison between Markdown and HTML:

Aspect	Markdown	HTML
Purpose	Designed for easy-to-read and write plain text formatting	Standard language for creating structured web pages
Syntax Complexity	Simple and minimalistic; easy to learn and use	More complex; requires understanding of various tags and attributes
Readability	Highly readable in raw form; closely resembles formatted output	Raw HTML can be cluttered and less readable
Functionality	Supports basic formatting (headings, lists, links, etc.); limited support for complex layouts	Offers comprehensive formatting capabilities, including complex layouts, multimedia embedding, and interactive elements
Use Cases	Ideal for writing documentation, README files, and content where simplicity and readability are prioritized	Essential for developing full-featured web pages and applications

Is ChatGPT response markdown?

Yes, ChatGPT uses markdown formatting in its responses. While ChatGPT generates Markdown-formatted text, the actual rendering—how the formatted text appears visually—is handled by the ChatGPT interface.