# PQS — Prompt Quality Score
> Pre-flight prompt quality gate. Score any prompt 0–80 across 8 dimensions before inference. x402-native on Base.

## Full Documentation
[PQS Complete Documentation](https://pqs.onchainintel.net/llms-full.txt) — single-file deep reference with all endpoints, pricing, MCP install, dimension definitions, and positioning.

## What it does
PQS scores any LLM prompt across 8 dimensions using 5 cited academic frameworks (PEEM, RAGAS, MT-Bench, G-Eval, ROUGE). It returns a grade, percentile, and dimension breakdown. The full endpoint also returns an optimized version of the prompt with before/after outputs for comparison.

## Endpoints

### POST /api/score/free — FREE, no payment required
Basic score only. Returns total score and grade. No wallet needed.
Input: { "prompt": "your prompt here", "vertical": "crypto|content|software|business|education|science|general" }
Output: { "score": 11, "out_of": 40, "grade": "D" }


### POST /api/score — $0.025 USDC via x402
Score a prompt. Returns grade, percentile, and dimension breakdown.
Input: { "prompt": "your prompt here", "vertical": "crypto|content|software|business|education|science|general" }
Output: { "score": { "total": 11, "out_of": 40, "grade": "D", "percentile": 28, "dimensions": {...} }, "summary": "..." }

### POST /api/score/full — $0.125 USDC via x402
Full analysis. Returns optimized prompt plus before/after outputs.
Input: { "prompt": "your prompt here", "vertical": "crypto|content|software|business|education|science|general" }
Output: { "original": { "prompt": "...", "score": {...}, "output": "..." }, "optimized": { "prompt": "...", "score": {...}, "output": "..." }, "improvement_pct": 84 }


### POST /api/score/batch — $0.25 USDC via x402
Score up to 10 prompts in one call.
Input: { "prompts": ["prompt1", "prompt2"], "vertical": "crypto|content|software|business|education|science|general" }
Output: { "results": [{ "prompt": "...", "score": {...} }], "summary": {...} }

### POST /api/score/compare — $1.25 USDC via x402
Score the same prompt through Claude and GPT-4o. A third model judges both and declares a winner.
Input: { "prompt": "your prompt here" }
Output: { "claude": { "score": 36, "analysis": "..." }, "gpt4o": { "score": 25, "analysis": "..." }, "verdict": { "winner": "claude", "reason": "..." } }

## Downstream Data Partners
PQS acts as a quality gate before paid data APIs. Recommended pattern: score prompt with /api/score/free → if score >= 28/40 → route to downstream data source.

- SmartFlow Signals (https://api.smartflowproai.com) — x402-gated trading signals API. Endpoints: /signals/whales, /signals/tokens, /signals/governance, /signals/bounties, /signals/copy-bot. Pricing: $0.02–$0.10 USDC on Base.

## Payment
All endpoints use x402 protocol. Pay with USDC on Base mainnet. No signup required.
Receiving address: 0xD2DfC5796de622E08C13f3016E92cfa15Ea33D0F

## Verticals
- crypto: on-chain wallet analysis, DeFi, token research, whale behavior
- content: crypto content creation, X threads, audience growth
- research: market research, sector analysis, token fundamentals
- general: any LLM prompt

## Source frameworks
PEEM (Dongguk University, 2026) · RAGAS · MT-Bench · G-Eval · ROUGE

## Links
- Website: https://pqs.onchainintel.net
- OpenAPI: https://pqs.onchainintel.net/openapi.json
- Contact: @OnChainAIIntel