Why AI Chatbot Text Looks Broken in WhatsApp — and How to Fix It

You copy a great answer from an AI chatbot, paste it into WhatsApp to share with a friend, a client, or a group — and it looks like a mess. Asterisks cling to every important word. Headings show their # symbols. A neat table turns into a pile of pipes and dashes. The text was perfect a moment ago. What happened?

This article explains the single underlying reason AI chatbot text breaks in WhatsApp — and the fastest reliable way to fix it, no matter which assistant you use.

The one-sentence explanation

AI chatbots write in Markdown; WhatsApp only supports a small, different set of formatting symbols. Everything else follows from that.

ChatGPT, Claude, Gemini, Microsoft Copilot, Perplexity, and almost every other assistant format their answers using Markdown — a widely used text-formatting language. Their apps render that Markdown into clean visuals, so on screen you see styled bold text and headings, never the raw symbols. But when you copy the text, you copy the symbols, not the visual styling.

WhatsApp has its own formatting system, and it is much smaller than Markdown. The pieces that overlap look fine; the pieces that do not overlap show up as raw characters.

Markdown vs WhatsApp, side by side

Feature	Markdown (AI output)	WhatsApp
Bold	**text**	*text*
Italic	*text* or _text_	_text_
Strikethrough	~~text~~	~text~
Headings	#, ##, ###	not supported
Tables	pipe tables	not supported
Inline code	`text`	`text`
Links	[label](url)	plain URL only

The mismatches in that table are exactly the things that look broken. Bold has the wrong number of asterisks. Headings and tables do not exist in WhatsApp at all. Links keep their Markdown brackets instead of becoming clickable. For a deeper treatment, see Markdown vs WhatsApp Formatting.

The five things that break most often

1. Bold becomes literal asterisks. **Important** shows as **Important** because WhatsApp expects a single asterisk.
2. Headings show their hashes. ## Summary stays ## Summary. WhatsApp has no headings, so the symbol just sits there.
3. Tables turn into pipe soup. Without table support, | Name | Price | becomes an unreadable row of characters.
4. Links keep their brackets. [our site](https://example.com) does not become a tidy link; both the label and the URL show.
5. Nested lists flatten. Indented sub-bullets lose their hierarchy.

Why you cannot just "turn it off" in the chatbot

People often look for a setting to disable Markdown. There usually is not one, because Markdown is how these models are trained to structure output — it is baked into their responses. You have two practical levers instead:

• Prompt for plain text. Ask the assistant to reply without Markdown and to use WhatsApp-style formatting. This reduces the problem but is not 100% reliable, especially in long answers.
• Convert the output. Run the reply through a tool that translates Markdown into WhatsApp formatting. This is the dependable option.

The reliable fix

Paste the chatbot's answer into the MD2WA editor. It reads the Markdown and produces a clean WhatsApp version:

• **bold** → *bold*
• Markdown italic → _italic_
• Headings → bold lines (the closest WhatsApp equivalent)
• Tables → readable bulleted lists
• Lists rebuilt with WhatsApp-friendly bullets

Then copy the result and send it. The whole process takes a few seconds and works for answers of any length — and it flags the 4,096-character limit if your message runs long.

A better prompt to start with

Adding a line like this to your request gets you closer to send-ready text:

"Reply in plain text formatted for WhatsApp: single asterisks for bold, underscores for italic, no headings, no tables, no Markdown links."

Treat it as a head start, not a guarantee. A final pass through the converter removes any leftover symbols before you hit send.

Per-assistant guides

The cause is identical across assistants, but if you want specifics:

• ChatGPT: How to Convert ChatGPT Replies to WhatsApp
• Claude and Gemini: Convert Claude & Gemini AI Answers to WhatsApp

Frequently asked questions

Is the chatbot formatting it wrong? No. The output is valid Markdown — it is just a different language from WhatsApp's formatting. You are translating between two systems, not fixing an error.

Will the person I message see asterisks too? If you paste raw Markdown, yes — they see the same broken symbols you do. Converting first means they see clean, intentional formatting.

Does this also affect Telegram, Slack, or Discord? Each app supports a different formatting subset. WhatsApp is one of the more limited ones, which is why AI output breaks there more visibly. The fix — convert to the target app's syntax — is the same idea.

Once you understand that this is a translation problem, it stops being frustrating. Translate the Markdown once with MD2WA, and your AI-written messages will look as sharp in WhatsApp as they did in the chat window.