How Is My PromptSign in

How is my prompt?

Find out before the AI does. Paste the brief you were about to send, and instead of an answer you get the read-back: a diagram of the task, the assumptions the AI would make, and every gap it would have to guess around.

YOUR BRIEF“Summarize Q3for the board”+ 1 spreadsheet attachedWITHOUT A READ-BACK?guesses silentlyWhich board?What tone?Which numbers?How long?FIRST DRAFTExpect Bad Response42/100SAME BRIEF · TWO OUTCOMESWITH HOW IS MY PROMPTShows its workq3-sales.csvJuly memo ✓RefineREFINED PROMPTReady to send98/100

Why it changes how you use AI

AI is only as good as the person managing it. Great results don't come from secret prompts — they come from briefing it well: goal, audience, constraints, source material. This tool makes that skill visible, so you learn to see how context actually flows into an answer.

Brief like a manager, not a genie-wisher

The people who get great results from AI aren't prompt wizards — they brief it like a delegate: goal, audience, constraints, source material. This tool makes that skill visible and repeatable.

Stop guessing what the model heard

When an answer misses, you rarely know why: ambiguous instruction? Forgotten attachment? Assumed context? The read-back isolates input quality from output quality.

The habit outlives the tool

After a few sessions you'll catch yourself pre-empting the gaps — stating the format, attaching the file, naming the audience — before any AI asks.

Stop burning budget on prompts you'll redo

A vague, low-effort prompt makes the AI guess — and you pay for every guess in tokens, then pay again to re-run it. Treating AI like a mind-reading magician is the most expensive way to use it. Seeing the gaps first means one good call instead of five wasteful ones.

How it works

Two passes, one loop: reflect, resolve, rewrite. It never answers the task — that restraint is the whole point.

feeds figures intowould add contextshapespasses tosets tone forproducesDATAq3-sales.csvMISSINGJuly memoMISSINGLast qtr updateAnalyzenumbersDraft summaryMatch toneOUTPUTBoard summary
Amber, dashed = referenced but never provided. That's what the AI would be guessing about.
Step 1

Paste the brief, attach the files

The exact prompt you were about to send to ChatGPT, Claude, or Gemini — PDFs, Word docs, spreadsheets, and images included.

Step 2

Read the read-back

One pass returns what a literal-minded new hire would understand: the task diagram, the assumptions it would make, and the specific gaps.

Step 3

Close the gaps, copy the result

Answer the wizard, add anything else that matters, and walk away with a refined prompt plus attach-this instructions.

What you get

A diagram of the task itself

Not a generic flowchart — your brief decomposed into its real steps: what feeds what, which operation the AI performs, and exactly where the holes are.

Gaps as a guided wizard

Every missing piece becomes one concrete yes/no question, answered one step at a time. Say yes and add the detail; say no and the tool designs around it.

A refined prompt you can ship

One click folds your answers into a rewritten, ready-to-paste prompt — plus a checklist of exactly what to attach when you send it.

A completeness score that means something

Every deduction maps to a specific, fixable sentence from your actual brief — never “consider adding more detail.”

Every session saved

Your read-backs and refined prompts live in your history with auto-generated titles, so the lesson compounds over time.

Share the read-back

One read-only link shows a teammate exactly what the AI heard — the fastest way to teach briefing habits without a lecture.

Pricing

A sharper prompt is worth more than it costs. A few dollars for answers that land the first time — instead of the hours (and tokens) you'd burn re-explaining yourself to an AI that guessed wrong.

$0 / forever
  • Limited daily usage
  • Every file type: PDF, Word, spreadsheets, images
  • Task diagrams, gap wizard, refined prompts
  • Full history and shareable links
Start free
$10 / month
  • Everything in Free
  • Unlimited analyses and refinements
  • Voice mode, with multi-model access
  • Direct access to the development team