Cost · Prompts

Prompt Explorer for Vague Requests

Paste a vague request. The page shows which of five writing rules are missing, then rewrites the text with those rules. It does not call a model. The Prompt Cost Simulator next door prices tokens. This page only cleans the words.

Paste a draft. An in-browser AI agent shows which writing checks are missing and rewrites the text. No model call. No API key. The Prompt Cost Simulator prices tokens. This page only cleans the words.

Five checks, said in plain words

Role. The text names who should answer, such as “you are” or “act as”.
Length. After filler words leave, the request is at least 40 characters.
Context. A line starts with “Context:” so the model knows which facts it may use.
Output. The text asks for a format, a list, a table, JSON, or bullets.
Limit. The text says what not to invent.

A worked example, before you paste anything

Starting text: “please make it good and write about our AI project thanks”.

Filler leaves. What remains: “And write about our AI project”.
That draft satisfies 0 of 5 checks. Clarity before the rewrite: 0.
The rewrite applies 8 notes, including the missing checks. Clarity after those checks: 100.
The box on the right fills with the same rewrite when the AI agent runs on this sample.

Clarity, rules, and the rewritten request

AI agent is ready. Run it, or edit the draft.

Clarity before — Checks already present, out of 100

Clarity after — After the missing checks are added

Notes applied — Filler removals plus missing checks

In plain words —

What changed

The AI agent lists these after it runs.

Rewritten request

Writing rules in this browser. Not a model answer, and not a claim that the rewrite is the best prompt.

How the rules work

Filler phrases are removed: please, make it good, be creative, thanks.
If no role is present, the page adds “You are a careful assistant.”.
If the cleaned draft is under 40 characters, it adds a line asking for a one-sentence task.
Missing context, output shape, and a do-not-invent limit are added as their own lines.
Clarity is how many of the five checks were already true, times 20. After the rewrite, those five are true, so the after-score is 100.

The AI agent fills this line after it runs.

Unlock a printable summary

The rewrite stays visible with no email. A work email unlocks Print / Save as PDF. Same subscribe flow as the newsletter.

FAQ

Does this call ChatGPT or another model?

No. The rewrite uses five writing rules in your browser: a role, a minimum length, a context line, an output shape, and a line that says not to invent facts. No API key. The draft is not sent.

How is this different from the Prompt Cost Simulator?

The Prompt Cost Simulator prices tokens and a monthly bill. This page only cleans the words. Use this page first if the request is vague, then price it on the other tool if you need a bill.

What does the clarity score mean?

Five checks. Each one that is already true is 20 points. A draft with none of them scores 0. After the rewrite those five are present, so the after-score is 100. The score is a checklist, not a grade from a model.

Do I need an API key?

No API key is required. No OpenAI, Anthropic, or Google key. The AI agent is a badge, a short thinking step, and a stamp. The rules run in your browser. The rewrite stays on screen before any email ask. Email is only for the printable summary.

Is the AI agent calling a cloud model?

No. The AI agent on this page is the in-browser analysis experience. It uses rule-based heuristics. The draft stays in the browser.

Will the rewrite always be a good prompt?

No. It adds the missing checks and removes a few filler phrases. You still have to put the real facts in the context line. The page does not know your product.