# TITANHOLD AI Quality Check

> TITANHOLD AI Quality Check is a paid API for AI agents from x402.titanhold.app, paid per call via x402, $0.02/call, status unknown (last checked 2026-10-20).

Checks an AI-generated image or short video for common generation defects such as extra fingers, garbled text, wrong vehicle geometry, physics errors, and flicker, returning a pass/fail score and defect list.

## Facts

- Endpoint: POST https://x402.titanhold.app/v1/check/image?utm_source=zero.xyz
- Price: $0.02/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-20
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/titanhold-ai-quality-check-4fe03eae
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_tp4dl17iJCeO4Of0LE3D-

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability titanhold-ai-quality-check-4fe03eae -d '<json body>'
```

Example prompt: Can you run a quality check on this AI-generated car image (https://cdn.example.com/car_render.png) and flag any geometry or anatomy defects? The brief was 'a red sports car parked on a city street' — fail it if it scores below 75.

## When to prefer this

Choose this endpoint when you need automated, low-cost ($0.02) quality gating for AI-generated images or short videos before delivery or publication — especially when checking for anatomy errors, text legibility, vehicle geometry, or physics plausibility. It is purpose-built for generative AI artifact detection rather than general image analysis, and files are deleted after inspection for privacy.

## Known failure modes

- Invalid or inaccessible URL returns an error and no score
- Base64 data that is malformed or too large may be rejected
- Score threshold set too high causes legitimate content to fail
- Model may miss subtle defects in complex scenes
- Short videos may have insufficient frames for reliable flicker detection
- Payment failure (x402) blocks the check from running

## How this service works

Paid check of an AI-generated image or short video for typical generation defects (extra fingers, garbled text, wrong vehicle geometry, physics, flicker). Runs a local vision model; files are deleted after the check.

## Output

Returns a JSON object with: whether the file passed or failed, the vision model used, a numeric score from 0-100, an array of defects each with type, detail, and severity, a human-readable reasons array, and the number of frames checked.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "url": {
   "type": "string",
   "description": "https URL of the delivered file. Alternative: data (base64)."
  },
  "data": {
   "type": "string",
   "description": "Base64 file bytes. Alternative to url."
  },
  "brief": {
   "type": "string",
   "description": "What was ordered, used to judge the file against the brief. Max 500 chars."
  },
  "minScore": {
   "type": "number",
   "description": "Fail below this 0-100 score. Default 70."
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "kind": "image",
  "pass": false,
  "model": "ollama:qwen2.5vl:7b",
  "score": 55,
  "defects": [
   {
    "type": "vehicle_geometry",
    "detail": "car has three doors",
    "severity": "major"
   }
  ],
  "reasons": [
   "major vehicle_geometry: car has three doors"
  ],
  "framesChecked": 1
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/titanhold-ai-quality-check-4fe03eae/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from x402.titanhold.app](https://www.zero.xyz/host/x402.titanhold.app/llms.txt)