MonkeybaseMonkeybase

Prompt workflow

Structure a Prompt

Turn a vague prompt into a reusable brief with a clear audience, task, constraints, and format.

Use this when: Clarify a prompt locally before sending it to an AI service.

Your input stays in the browser — nothing is sent to a server. Privacy policy.

Start with this messy input

A request with no audience or deliverable

Could you maybe help write an update about our launch? Make it good and not too long.

Steps

Run 4 steps.

  1. 1

    Prompt Mirror

    Expose missing structure

    Paste the rough request and add role, task, output format, and constraints.

    Look for: Missing audience, evidence boundaries, and a concrete deliverable.

    Open Prompt Mirror in a new tab
  2. 2

    Text Workbench

    Clean the instruction text

    Normalize the prompt template and remove conversational hedging.

    Look for: A prompt that can be reused without editing its intent each time.

    Open Text Workbench in a new tab
  3. 3

    Readability Checker

    Check clarity for collaborators

    Read the template as instructions another teammate must follow.

    Look for: Long constraints or vague words that could cause different interpretations.

    Open Readability Checker in a new tab
  4. 4

    Diff Viewer

    Preserve the improvement

    Compare the original ask and the reusable template.

    Look for: Which assumptions became explicit and which still need your judgment.

    Open Diff Viewer in a new tab

Before

Could you maybe help write an update about our launch? Make it good and not too long.

A reusable prompt template

Role: Product update editor.
Task: Draft a launch update for [feature].
Audience: Existing users who need to decide whether to try it.
Constraints: Use only verified claims; 90 words maximum; no urgency language.
Format: Headline, two-sentence explanation, and one CTA.

What the tools cannot decide for you

  • A clearer prompt does not ensure accurate output from a model.
  • You must supply real product facts, approved tone, and any required safety boundaries.
  • Review generated output before using it in a launch or customer communication.

Try your own input

Start in Prompt Mirror

Reveal which parts of a rough request need structure.

Open the first tool

Continue from here

Prompt Mirror Patterns

Read reusable patterns for clearer prompts.

Prompt Mirror vs Manual Editing

Learn where structure helps and judgment still matters.

Prompt Mirror

Open the local rules-based prompt experiment.

Found an issue or have a suggestion? Report an issue or suggest an improvement.