Prompt Score
Know how good a prompt is before you run it
Prompt Score rates every prompt you save from 1 to 5 against 6 weighted criteria, and tells you exactly what to fix. No more guessing why one prompt works and another returns noise.
Prompt Score
Live exampleWrite a post about our product launch.
The rubric
Six weighted criteria, one honest number
Every prompt is evaluated against the six things that actually separate a strong prompt from a weak one. Each criterion returns its own assessment and a targeted tip; the weighted average is your score.
Clarity & Specificity
25%Is the task unambiguous? Precise verbs, defined output, no room for interpretation.
Context
20%Does the model get the background it needs: situation, constraints, boundaries?
TCOF Structure
20%Task, Context, Output, Format: is the prompt organized so nothing gets lost?
Role Prompting
15%Does the prompt assign the model a clear role or persona to reason from?
Chain-of-Thought
10%Is the model guided to reason through the problem where the task needs it?
Few-shot Examples
10%Are there concrete examples showing the expected input and output?
How it works
From saved prompt to score in seconds
Save a prompt
Paste any prompt into your library: ChatGPT, Claude, Gemini, it does not matter. The score evaluates the prompt, not the model.
Get the score and the why
The prompt is rated 1 to 5 on the six criteria, each with its own assessment, so you see exactly where it loses points.
Fix, re-score, keep the best
Apply the targeted tips (or run Quick Optimize), score the new version, and keep the one that measurably wins. Every version keeps its score.
Why it matters
A score turns prompt quality from a feeling into a decision
- Stop re-running weak prompts: know which ones need work before they waste a call.
- Compare two versions with a number, not a hunch, and keep the one that scores higher.
- Onboard teammates onto a shared library where quality is visible at a glance.
Frequently asked questions
What is a prompt score?+
A prompt score is a 1-to-5 rating of how well an AI prompt is written: how clear, specific, and structured it is, independent of any single model. Keep My Prompts scores every prompt you save so you know which ones are strong and which need work before you run them.
How is the Prompt Score calculated?+
Each prompt is evaluated against 6 weighted criteria: Clarity & Specificity (25%), Context (20%), TCOF Structure (20%), Role Prompting (15%), Chain-of-Thought (10%), and Few-shot Examples (10%). The weighted average becomes the 1-to-5 score, and each criterion comes with a targeted tip.
Is Prompt Score free?+
Yes. Prompt Score is included on every plan, including the Free plan. You can score every prompt you save without any usage limit.
Which AI models does it work with?+
All of them. The score evaluates the prompt itself, not a specific model, so it works whether you run the prompt on ChatGPT, Claude, Gemini, or any other AI assistant.
Can I track score improvements over time?+
Yes. Every prompt version keeps its own score, so when you rewrite a prompt you can compare versions side by side and keep the one that actually scores higher.
Go deeper
How we think about prompt scoring
Score your first prompt in the next two minutes
Free plan, no card. Save a prompt, get its score, and know exactly what to improve.
Get started free