Independent extraction guides · Built for curious developersSource → Schema → Something useful
Documentation / Start here

One source.
One useful record.

Begin with a small, explicit contract. Inspect the example below, download the files, and adapt the field meanings to your own authorized workflow.

STATIC EXAMPLES · NO CREDENTIALS REQUIRED
The first handoff

Make absence
part of the design.

This example describes a fictional catalog item. The price is represented as decimal text, the source is identified, and unavailable availability remains null. The issue list explains the absence.

The timestamp belongs to the illustration. It is not an observation from a live source. These downloadable files do not submit requests, authenticate a user, or connect to an extraction service.

Get the example files
sample-record.json
{
  "schema_revision": "1.0",
  "source_id": "catalog-demo-01",
  "observed_at": "2026-09-13T00:00:00Z",
  "record": {
    "name": "Example reusable bottle",
    "price_amount": "24.00",
    "currency": "USD",
    "availability": null
  },
  "issues": [
    "availability_not_present"
  ]
}
Illustrative data · Explicit nulls · A contract, not a live API response

Read the contract, not just the keys

The envelope identifies the source and observation context. The record contains selected business fields. The issue list preserves an explicit reason for a missing value. Use a different representation when your consumer needs one, but document the decision rather than leaving it to interpretation.

Example fieldMeaning in this contract
schema_revisionA visible version for the record contract.
source_idA stable identity for the approved source.
observed_atWhen the source was observed, distinct from publication time.
record.nameThe item this record describes.
record.price_amountDecimal text chosen for this example; not a numeric identifier.
record.currencyCurrency context for the price.
record.availabilityNull because the example supplies no supported availability value.
issuesA machine-readable explanation of missing or rejected fields.

Four checks before delivery

Parse: confirm the output is a valid serialization. Validate structure: check types, required fields, and unexpected properties. Check meaning: apply the rules required by the task. Check evidence: confirm that important values are supported by the intended source item.

The supplied schema is a structural example. It does not prove that a source exists or that the values are true. Date-time format checking also depends on the validator and its configuration. Keep these checks separate in your implementation and include expected failures in the review set.

Primary reference
JSON Schema · object reference ↗

Consult the source documentation for implementation details and current access conditions. The workflow above is an editorial planning pattern.

Download the examples

Connect the pattern to an implementation

Choose the source and access route, then implement retrieval and extraction in an environment appropriate to the workload. Keep credentials out of static pages and client bundles. Add output validation before connecting records to a downstream action.

Start with the data contract article, continue to the JSON validation guide, and use the deployment overview when selecting a runtime. None of the examples require or imply an ExtractAPI.com API key.

Make messy data
your next good idea.

Find your starting point