PQS/for agents
x402 native·Base + Solana
Prompt Quality Score

Pre-flight checks for agent workflows

PQS scores AI prompts 0–80 across 8 quality dimensions before they hit expensive LLM endpoints. Catch weak prompts early. Reduce wasted inference spend. Build more reliable agents.

Pay per call · No subscription · One email to start
Generate an API key and start scoring in 2 minutes.
Get API key →
The Problem

Bad prompts reach your LLM before you know they’re bad

Silent agent failures
Your agent sends a vague prompt, the LLM hallucinates or returns garbage, the downstream tool breaks. No error thrown. No alert. Just a bad result your user sees.
Wasted inference spend
Every weak prompt still costs you a full API call. At scale, that's real money spent on outputs your agent can't use.
PQS is the quality gate
Score the prompt before it hits the LLM. If it scores below your threshold, reject it, rewrite it, or route it differently. $0.025 per check vs $0.01-$0.10 per wasted inference call.
How It Works

Four lines. One quality gate.

1
Agent has a prompt
Your agent constructs a prompt for an LLM call.
2
POST to PQS
Score comes back in <500ms with grade, dimensions, and fixes.
3
Gate on score
If score >= 28, proceed. Below threshold? Rewrite or reject.
4
Pay $0.025 via x402
Automatic micropayment on Base or Solana. No API key needed for x402.
New: Preflight

Prompt quality + endpoint trust in one call

Pass/Fail before you waste budget on a doomed call. Decide whether to make a paid LLM call before you make it. Preflight scores your prompt against an 8-dimension quality rubric (pass/fail at 28/80) and optionally pulls the target endpoint's on-chain trust signals — total payments, unique payers, last activity. Two-in-one circuit breaker for agents that don't want to waste budget on weak prompts or sketchy endpoints. Built on 6,511 verified endpoints mapped from Base mainnet facilitator data.

0-80
Prompt score
0-1.0
Trust score
3 levels
Clearance
$0.05
Price
POST /api/preflight
{ "prompt": "...", "endpoint_url": "https://mesh.heurist.xyz/..." }

→ "clearance": "APPROVED", confidence: 0.83
API Example

Score any prompt in one call

Request
POST /api/score HTTP/1.1
Host: pqs.onchainintel.net
Content-Type: application/json

{
  "prompt": "Analyze the mass spectral data for compound X",
  "vertical": "science"
}
Response
{
  "pqs_version": "2.0",
  "score": {
    "total": 31,
    "out_of": 80,
    "grade": "C",
    "percentile": 39,
    "dimensions": {
      "clarity": 5,
      "specificity": 3,
      "context": 4,
      "constraints": 3,
      "output_format": 2,
      "role_definition": 5,
      "examples": 4,
      "cot_structure": 5
    }
  },
  "top_fixes": [
    "Specify the instrument type and ionization method",
    "Define the expected molecular weight range",
    "Request a structured output format for the analysis"
  ]
}
6,511
verified x402 endpoints
630+
new endpoints per week
$0.025
per score via x402
Heurist Mesh · Coming Soon
PQS is being integrated as a native Mesh agent
The Mesh endpoint is built and ready on our side. Once live, agents on Heurist Mesh will get 100 free calls per day with no payment infrastructure needed.
Learn about Mesh →
Pricing

Free tier or pay per call

Free
Heurist Mesh
100 calls / day · coming soon
  • ✓Score + grade + verdict
  • ✓General vertical only
  • ✓No API key required
  • ✓No payment infrastructure needed
Paid
Direct API
$0.025 / call via x402 or API key
  • ✓Full 8-dimension breakdown
  • ✓All 8 verticals
  • ✓Preflight clearance ($0.05)
  • ✓Prompt optimization ($0.125)
  • ✓Batch scoring ($0.25)
  • ✓Cross-model comparison ($1.25)
  • ✓All 5 academic frameworks
  • ✓Base + Solana networks

Start scoring in 2 minutes

Generate an API key, POST a prompt, get a score. Email-based key. No contracts. No subscription.

PQS by OnChainIntel · pqs.onchainintel.net