What is a Claude prompt enhancer?

A Claude prompt enhancer is a tool that rewrites a rough draft into the structure Claude follows best — role, context, task and output format, expressed in Anthropic's XML tag style. Prompt Enhancer does this from a global hotkey, in place, in any app, in under four seconds.

Does it work in Claude.ai and Claude Code?

Yes. Prompt Enhancer is system-wide, so it works in Claude.ai in the browser, in Claude Code, in Cursor, in your editor, Notion, Slack — anywhere there's a text field. You select your draft, press the hotkey, and the structured prompt replaces it on the spot.

There are 3 free enhancements with no signup, then a 7-day free trial (email unlock, no password). BYOK — your own Anthropic key, direct to api.anthropic.com — is a Pro feature. Pro Lifetime is $29 one-time, no subscription.

Why XML for Claude specifically?

Claude was trained on a large amount of XML-tagged content and parses tagged sections as labeled containers. Anthropic's own docs name XML the primary recommended method for structuring complex prompts. That's why Prompt Enhancer outputs role/context/task/output_format tags rather than a longer paragraph.

Claude Prompt Enhancer

Name: Prompt Enhancer
Author: Prompt Enhancer

Prompt Enhancer is a desktop Claude prompt enhancer for macOS and Windows. Select any rough draft, press one global hotkey (⌃⌥⌘P on Mac, Ctrl+Alt+P on Windows), and it rewrites your text into the structure Claude follows best — <role>, <context>, <task>, <output_format> in Anthropic’s official XML format — in place, in under four seconds.

No window switching. No copy-paste. The cursor stays exactly where it was.

Why structure matters for Claude

Claude is not GPT. It was trained on a large amount of XML-tagged content, so it parses tagged sections as labeled containers instead of one undifferentiated paragraph. Anthropic’s documentation names XML the primary recommended method for structuring complex prompts.

A vague paragraph makes Claude guess. A structured prompt tells it exactly who to be, what context to use, what the task is, and how the answer should look — so you stop re-prompting.

For the full breakdown of the format, read How to structure Claude prompts with XML.

How it works

Write a rough draft anywhere — Claude.ai, Claude Code, Cursor, VS Code, Notion, Slack.
Select it and press the hotkey.
In under four seconds the selection is replaced with a structured Claude prompt:

<role>You are a senior backend engineer reviewing a pull request.</role>
<context>Node.js REST API, TypeScript, PCI-DSS compliance required.</context>
<task>Review the diff for security vulnerabilities.</task>
<output_format>Numbered list: severity, file/line, explanation, fix.</output_format>

It works system-wide — macOS uses the Accessibility API, Windows uses UI Automation, Electron/web apps fall back to clipboard automatically.

Built for the way you actually use Claude

Prompt Enhancer ships target-tool presets, so the output matches where the prompt is going. The two for Claude:

Chat — for Claude.ai and general chat: adds role, constraints and output format as clean XML.
Claude Code — for Claude Code and Cursor: structures coding requests with language/framework, expected output, and test expectations.

(Other presets cover Codex, image models, video, and deep research — one app, the right structure per target.)

Privacy: bring your own key

In BYOK mode your Anthropic API key is stored in the macOS Keychain / Windows Credential Manager, and traffic goes directly to api.anthropic.com — no proxy, no server, no logs. You pay Anthropic directly; we take no cut.

Honest caveats

The macOS DMG is Apple-notarized since June 2026 — installs without Gatekeeper warnings. The Windows installer may show a one-time SmartScreen prompt.
Public beta launched 2026-05-23, v0.1.1 on 2026-05-27 — no long track record yet.
Enhancement needs an internet connection (it calls the Claude API).

See pricing and download →