Prompt Quality and Suggestions
The Prompt Quality Engine helps users write more effective prompts by evaluating their input and offering dynamic suggestions in real time. This improves the accuracy, relevance, and performance of AI-generated smart contracts or web builds.
🧠What It Does
Analyzes your prompt as you type
Scores the quality based on clarity, completeness, and actionability
Suggests improvements or edits
Offers quick-fill prompt templates to help you get started
🔍 How Prompt Scoring Works
Every prompt entered into the platform is evaluated based on the following:
Clarity
Does the prompt clearly define what’s being built?
Structure
Are key elements like name, supply, logic, etc. present?
Feasibility
Can the AI actually perform what’s being asked?
Specificity
Is the behavior or logic unambiguous?
Prompts are given a score from 0 to 100, displayed below the input box with a confidence meter and optional tips.
Smart Suggestions
If the score is low, the system will:
Highlight missing components (e.g., no lockup period defined in staking prompt)
Suggest improvements (e.g., “Consider defining token name and reward structure”)
Offer “Improve Prompt” button that rewrites your input using AI
Prompt Templates & Examples
The prompt system also includes:
Example prompts that update based on the selected use case (token, web, AI, etc.)
One-click fill buttons to inject a prompt starter into the input
“Remix” Feature (if enabled): lets users click to remix their current prompt into 3–5 smart variations
Sample Low vs High Quality Prompt
Low
“Make token”
High
“Create a staking pool token with 12-month vesting, 9% APR, and a 15-day lockup. Token name: STAKEGEN.”
UX/UI Placement
Located directly beneath the main prompt input field
Displays score, guidance, and suggestions in a collapsible panel
Suggestions update live as the prompt changes
Limitations
AI suggestions are non-binding — they won’t block generation
Scores are heuristic, not definitive
Overly creative or abstract prompts may not score high but could still produce useful outputs
Last updated