publishedv1

Manufacturing Defect Locator API

Send a photo URL and get the bounding box of the manufacturing defect on the part.

Overview

The Manufacturing Defect Locator API detects manufacturing defects in product photos and returns their location as a fractional bounding box. Send a POST request with an image_url to the Detect Defect endpoint, and receive a bounding_box with normalized coordinates, a found boolean indicating whether a defect was detected, and a confidence score. It is built for quality assurance teams, manufacturing engineers, and developers building inspection or automation tools who need a fast, reliable way to flag defective parts without building custom vision pipelines.

Endpoint 1

Detect Defect

Find a manufacturing defect in an image and return its location as a fractional bounding box, a flag indicating if a defect was found, and a confidence score.

activev1json
POST
https://api.init.computer/v1/428tjnocuf
Request body1 field
image_urlurlURL of the image to analyze
Response3 fields
bounding_boxobject

Fractional coordinates of the defect bounding box

topnumber

Top coordinate as a fraction of image height

leftnumber

Left coordinate as a fraction of image width

bottomnumber

Bottom coordinate as a fraction of image height

rightnumber

Right coordinate as a fraction of image width

foundboolean

Whether a defect was detected in the image

confidencenumber

Confidence score for the detection (0‑1)

Try for Free
Examplesrequest & response
Request
curl -X POST https://api.init.computer/v1/428tjnocuf \
  -H "Authorization: Bearer $INIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"image_url":"https://…"}'
Response
{
  "bounding_box": {
    "top": 0,
    "left": 0,
    "bottom": 0,
    "right": 0
  },
  "found": true,
  "confidence": 0
}
View raw init.json
{
  "tables": [],
  "project": {
    "id": "d93b34cc-27c6-41e6-b960-04af39777b04",
    "name": "Manufacturing Defect Locator",
    "version": 1,
    "description": "Send a photo URL and get the bounding box of the manufacturing defect on the part."
  },
  "resources": [
    {
      "id": "862350c3-81a1-4fd7-ba5b-8b425768fb60",
      "data": {
        "create": {
          "tables": [],
          "buckets": []
        },
        "tables": [],
        "buckets": []
      },
      "name": "Detect Defect",
      "input": {
        "input": [
          {
            "name": "image_url",
            "type": "url",
            "fields": [],
            "required": false,
            "description": "URL of the image to analyze"
          }
        ],
        "description": "Input for Detect Defect"
      },
      "route": "https://api.init.computer/v1/428tjnocuf",
      "method": "POST",
      "output": {
        "output": [
          {
            "name": "bounding_box",
            "type": "object",
            "fields": [
              {
                "name": "top",
                "type": "number",
                "fields": [],
                "required": false,
                "description": "Top coordinate as a fraction of image height"
              },
              {
                "name": "left",
                "type": "number",
                "fields": [],
                "required": false,
                "description": "Left coordinate as a fraction of image width"
              },
              {
                "name": "bottom",
                "type": "number",
                "fields": [],
                "required": false,
                "description": "Bottom coordinate as a fraction of image height"
              },
              {
                "name": "right",
                "type": "number",
                "fields": [],
                "required": false,
                "description": "Right coordinate as a fraction of image width"
              }
            ],
            "required": false,
            "description": "Fractional coordinates of the defect bounding box"
          },
          {
            "name": "found",
            "type": "boolean",
            "fields": [],
            "required": false,
            "description": "Whether a defect was detected in the image"
          },
          {
            "name": "confidence",
            "type": "number",
            "fields": [],
            "required": false,
            "description": "Confidence score for the detection (0‑1)"
          }
        ],
        "description": "Output from Detect Defect"
      },
      "status": "active",
      "version": 1,
      "description": "Find a manufacturing defect in an image and return its location as a fractional bounding box, a flag indicating if a defect was found, and a confidence score.",
      "requestFormat": "json"
    }
  ],
  "schemaVersion": 1
}

Use cases

How teams put the Manufacturing Defect Locator API to work.

01

Automated Line Inspection

Manufacturing engineers use Detect Defect to check camera images captured at production stations. They build inspection dashboards that display each part, mark the returned fractional bounding box, and flag images where found is true for rejection or manual review.

02

Quality Review Queues

Quality teams build a review queue that groups part photos by defect status and confidence. Detect Defect supplies the found flag, confidence score, and bounding box, helping inspectors prioritize uncertain results and open each image directly to the suspected defect.

03

Supplier Intake Screening

Supplier quality managers add photo-based screening to incoming-part workflows. Their receiving app sends each inspection image to Detect Defect, then routes flagged parts—with the marked defect location and confidence—to an inspector before the shipment is accepted.

04

Production Traceability Records

Factory software developers attach visual inspection results to part and batch records. Detect Defect returns the bounding box, found status, and confidence for each image, so teams can retain defect evidence, compare inspection outcomes, and investigate recurring production issues.

Questions

How do I get manufacturing defect location data as JSON?

Call Init Computer's Manufacturing Defect Locator API. POST a JSON body with image_url to the Detect Defect endpoint at https://api.init.computer/v1/428tjnocuf and you'll get back bounding_box (fractional coordinates), found (boolean), and confidence (score). Use the CLI: init run https://api.init.computer/v1/428tjnocuf --data image_url=https://example.com/part.jpg

Is there an API for detecting defects in product photos?

Yes, Init Computer's Manufacturing Defect Locator API detects defects in part photos. The Detect Defect endpoint takes an image_url and returns a fractional bounding_box, a found flag, and a confidence score. It's a ready-to-run backend API you can call with an API key, fork in the console, or invoke from the CLI or JavaScript SDK.

What's the best way to flag defective parts without manual inspection?

Use the Manufacturing Defect Locator API to automate the first pass. Send each part photo to the Detect Defect endpoint (POST https://api.init.computer/v1/428tjnocuf with image_url) and check the found field. If found is true and confidence is high, route the part to rework or rejection. This gives you a consistent, repeatable QC check.

Can I call the defect locator from Python or Node?

Yes. In Node, use the @init-computer/sdk: createClient() then init.run('https://api.init.computer/v1/428tjnocuf', { image_url }). In Python or any language, POST JSON to the same route URL with an Authorization: Bearer $INIT_API_KEY header. The response includes bounding_box, found, and confidence.

How much does the Manufacturing Defect Locator API cost?

Each call to the Detect Defect endpoint is billed as one run to your Init Computer account. There are no per-image or per-pixel fees beyond that. You pay per API call, and you can see usage in your Init console. The endpoint returns bounding_box, found, and confidence for each image you send.

Call from the CLI

Run the Manufacturing Defect Locator API from your terminal with any Init API key. Runs are billed to your account.

npm install -g @init-computer/cli
init login --api-key init_sk_…

01Detect Defect

init run /v1/428tjnocuf \
  --data image_url=https://example.com

Inputsimage_urlurl

Returnsbounding_boxobjectfoundbooleanconfidencenumber

Call from the SDK

Call the same endpoints from Node.js. Keep INIT_API_KEY server-side.

npm install @init-computer/sdk

01Detect Defect

import { createClient } from "@init-computer/sdk";

const init = createClient(); // reads INIT_API_KEY
const result = await init.run("https://api.init.computer/v1/428tjnocuf", {
  "image_url": "https://example.com"
});
console.log(result);

Inputsimage_urlurl

Returnsbounding_boxobjectfoundbooleanconfidencenumber

More public APIs developers pair with this one.

View all