Paste any AI prompt below and get an instant quality score across 6 dimensions. No API key needed — scoring is heuristic and runs in your browser.
Try scoring one of these:
Your prompt is evaluated across 6 dimensions: Role & Persona (does it define who the AI should be?), Specificity (concrete numbers vs vague words), Output Format (does it specify JSON, list, table, etc.?), Constraints (does it say what NOT to do?), Context (does it explain the goal?), and Examples (does it show the expected style?). Each category has a max score; the total is out of 100.
Higher-scoring prompts get more consistent, usable outputs from AI models. A prompt scoring below 50 typically produces vague or off-format responses. One above 80 gives the model a clear role, concrete goal, expected format, and guardrails — which correlates with much higher first-try success rates.
No. The prompt scorer is entirely heuristic-based and runs in your browser with no API calls. It's instant and free. If you want an AI to actually rewrite your prompt for you, use the main Prompt Improver tool (which does require your own Anthropic API key).
A score of 70+ (grade B) is solid for most everyday tasks. A score of 85+ (grade A) is ideal for complex structured tasks like data extraction, API integrations, or multi-step reasoning. Scores below 50 usually mean the prompt is missing at least two of: a role, a format spec, or concrete constraints.
The scorer shows specific tips for each category. The highest-ROI fixes are: (1) Add "You are an expert [domain]…" at the start. (2) Specify the output format ("Return a JSON object with keys:…"). (3) Add a negative constraint ("Do not include preamble"). (4) Replace vague words like "good" or "appropriate" with concrete criteria. Use the Prompt Improver tool on this site to have Claude do all of this automatically.