Format AI Chat Prompts — Free Prompt Formatter for GPT, Claude, Gemini

Structuring multi-turn conversations for AI APIs isn’t always straightforward — especially when you need to visualise the flow, copy formatted JSON, or prepare prompts for multiple LLMs. SimpleTools Prompt Formatter makes it easy to build, visualise, and export AI chat prompts — all in your browser.

What Is a Multi-Turn Chat Prompt?

Most modern LLM APIs (OpenAI, Anthropic, Google Gemini, Cohere) accept prompts as a structured list of messages, each with a role and content:

[
  {"role": "system", "content": "You are a helpful coding assistant."},
  {"role": "user", "content": "How do I reverse a string in Python?"},
  {"role": "assistant", "content": "You can reverse a string using slicing: s[::-1]"},
  {"role": "user", "content": "What about in JavaScript?"}
]

This format is called the chat completions format and is used by:

• OpenAI: gpt-4o, gpt-4-turbo, etc.
• Anthropic Claude: claude-3-5-sonnet, claude-3-opus
• Google Gemini: via the Gemini API with contents format
• Cohere, Mistral, LLaMA (via local servers or cloud APIs)

What the Prompt Formatter Does

• Add messages with system, user, or assistant roles
• Visual conversation view — see the prompt as a chat timeline, not raw JSON
• Reorder and delete messages with drag-and-drop
• Export as JSON — the raw messages array ready to paste into API calls
• Export as formatted text — for sharing or documentation
• Copy individual messages or the full conversation
• Character and token count — see how large your prompt is

Why Build Prompts in the Browser?

Your prompts often contain sensitive business logic — system instructions that define your AI product’s behaviour, proprietary persona definitions, confidential use case instructions. These shouldn’t be sent to third-party services.

SimpleTools Prompt Formatter:

✅ Your prompts never leave your browser tab
✅ Works offline — build and format prompts without internet
✅ No account or API key needed
✅ Free for unlimited prompts and messages

How It Works

The tool maintains an in-memory array of message objects. The UI components for adding, editing, reordering, and removing messages are all JavaScript operating on that array. When you export:

• JSON export: JSON.stringify(messages, null, 2) produces the formatted JSON
• Text export: A template literal formats each message as a labelled block

There’s no backend, no database, no API. All state lives in your browser tab’s JavaScript runtime.

How to Use the Prompt Formatter

1. Visit simpletools.one/prompt-formatter
2. Click + Add System Message to add your system prompt
3. Click + Add User Message to add the first user turn
4. Click + Add Assistant Message to add an example assistant response (for few-shot prompting)
5. Continue adding turns as needed
6. Click Copy JSON to get the API-ready messages array
7. Click Export to download as a .json file

Prompt Engineering Tips

System message best practices:

• Define the assistant’s role, persona, and constraints clearly
• Specify the output format if you need structured responses
• Include guardrails for what the assistant should and shouldn’t do

Few-shot examples:

• Add user/assistant pairs before the final user message to show the model the expected response format
• Keep examples relevant to the task — unrelated examples can confuse the model

Context management:

• LLMs have context limits (measured in tokens) — use the token counter to stay within limits
• Prioritise the most recent context; older turns can be summarised or removed

Use Cases

• Prompt engineering: Build and iterate on prompts before integrating them into code
• Documentation: Format and share example conversations for technical documentation
• Few-shot learning: Construct few-shot prompts with example user/assistant pairs
• Testing: Prepare test cases for LLM evaluation runs
• Teaching: Create structured examples for AI education content

Format your AI prompts at simpletools.one/prompt-formatter — private, structured, and completely free.