AI workflows

Structured ecommerce data your agent can use.

Start with a public product, collection or store URL. datawebot. creates a durable, bounded request and returns source-reported records through the same API used by the browser.

Synthetic catalog data · not a retrieval result
02 / Inspect the example
Synthetic example JSON · no source data
{
  "example_type": "synthetic",
  "notice": "Every product name, identifier, option, price and availability value is invented for illustration. This is not a datawebot retrieval, a merchant catalog or a live offer.",
  "products": [
    {
      "id": "example-product-001",
      "title": "Cotton Crewneck T-Shirt",
      "url": null,
      "vendor": "Example Outfitters",
      "product_type": "Tops",
      "description_html": "Fictional example record for demonstrating product and variant fields.",
      "images": [],
      "variants": [
        {
          "id": "example-variant-001",
          "title": "Small / Navy",
          "sku": "DEMO-TEE-S-NV",
          "options": {
            "Size": "Small",
            "Color": "Navy"
          },
          "price": {
            "amount": "32.00",
            "currency": "USD"
          },
          "compare_at_price": {
            "amount": "38.00",
            "currency": "USD"
          },
          "available": true
        },
        {
          "id": "example-variant-002",
          "title": "Medium / Navy",
          "sku": "DEMO-TEE-M-NV",
          "options": {
            "Size": "Medium",
            "Color": "Navy"
          },
          "price": {
            "amount": "32.00",
            "currency": "USD"
          },
          "compare_at_price": {
            "amount": "38.00",
            "currency": "USD"
          },
          "available": true
        },
        {
          "id": "example-variant-003",
          "title": "Medium / Sand",
          "sku": "DEMO-TEE-M-SD",
          "options": {
            "Size": "Medium",
            "Color": "Sand"
          },
          "price": {
            "amount": "34.00",
            "currency": "USD"
          },
          "compare_at_price": null,
          "available": false
        },
        {
          "id": "example-variant-004",
          "title": "Large / Navy",
          "sku": "DEMO-TEE-L-NV",
          "options": {
            "Size": "Large",
            "Color": "Navy"
          },
          "price": {
            "amount": "32.00",
            "currency": "USD"
          },
          "compare_at_price": null,
          "available": null
        }
      ],
      "provenance": {
        "storefront_id": null,
        "source_url": null,
        "retrieved_at": null,
        "schema_version": "illustrative"
      },
      "warnings": [
        "Synthetic example only; not collected from a source."
      ]
    }
  ],
  "warnings": [],
  "errors": [],
  "completeness": "illustrative_example",
  "pages_fetched": null,
  "credits_used": null,
  "billing_model": "not_applicable",
  "visits": []
}
Faithful extraction

Retrieve facts. Keep decisions in your agent.

datawebot. preserves products and variants without inventing missing attributes, recommendations or confidence scores. Your workflow decides how to compare, enrich or analyze the result.

Review the data contract
Products

Source IDs, titles, vendors, descriptions and image links where provided.

Variants

Variant IDs, SKUs, named options, prices and observed availability.

Provenance

Source URL, storefront identity, retrieval time, schema version and warnings.

Boundaries

Explicit page and credit ceilings keep autonomous requests within owner authority.

Designed for asynchronous agents.

Submit once, persist the extraction ID, and continue when the saved request changes state.

01

Discover

Read capabilities, normalize the URL and estimate the maximum cost.

02

Run

Submit with an idempotency key and an owner-authorized credit ceiling.

03

Retrieve

Honor the polling hint, collect every result page and inspect completeness.

A few useful answers.

Connect an agent.

Create a constrained API key, then follow the complete execution loop.

Create an agent key