Penny Engine: JSON Checks & Catalog CSV Audit avatar

Penny Engine: JSON Checks & Catalog CSV Audit

Pricing

$50.00 / 1,000 completed work orders

Go to Apify Store
Penny Engine: JSON Checks & Catalog CSV Audit

Penny Engine: JSON Checks & Catalog CSV Audit

Check up to 20 JSON results against your rules, or audit up to 50 catalog rows for missing fields, duplicate SKUs and margin arithmetic. Get JSON and Markdown issue reports. $0.05 per completed job. Checks supplied data; does not fetch data or verify facts.

Pricing

$50.00 / 1,000 completed work orders

Rating

0.0

(0)

Developer

kasheik mcgee

kasheik mcgee

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

a day ago

Last modified

Categories

Share

Penny Engine · Agent QA and Catalog Quality Reports

Buy a completed, deterministic digital work order. This Apify Actor is designed to give agents and automation operators a machine-readable JSON quality report plus a Markdown summary, without a separate x402 wallet or manual invoice. Apify handles the billing when you configure the Actor's pay-per-event price.

Supported jobs

  1. agent-output-batch-qa — check 1–20 results from an agent workflow against user-supplied assertions: required paths, nonempty paths, expected values, and numeric ranges. Each item returns PASS/FAIL, failed paths, and an input-content SHA-256 fingerprint.
  2. catalog-csv-audit — audit 1–50 catalog products provided as CSV text (sku, name, optional price, cost, shippingCost, and feePct). Return duplicate SKU flags, missing required fields, and estimated contribution and margins. These are arithmetic estimates, not validated market prices.

The dataset receives exactly one output record per successful work order, containing deliverables.reportJson, deliverables.reportMarkdown, and evidence.sha256.

Example 1 — agent result QA

{
"jobType": "agent-output-batch-qa",
"payload": {
"jobs": [
{"id": "step_1", "result": {"status": "done", "records": 5}, "expectations": {"equals": {"status": "done"}, "numberRanges": {"records": {"min": 1}}}}
]
}
}

Example 2 — catalog quality

{
"jobType": "catalog-csv-audit",
"payload": {"csv": "sku,name,price,cost,shippingCost,feePct\nA1,Widget,24,10,3,10\nA2,Other,19,12,2,8", "targetMarginPct": 25}
}

Pricing, refunds and limits

This Actor uses Actor.pushData(result, "work-order-completed") to save the result and bill the one named event together after successful computation. You must configure the price for that event in Apify Console, e.g. an introductory $0.05 per completed work order, and test its paid-tier budget behavior. No completed event is charged if input validation fails. The SDK enforces the run budget; a charge result with chargedCount other than one is rejected. Users on Apify free plans do not generate developer revenue under the current platform model. Apify's documented profit formula is 0.8 × event revenue − platform costs; payouts are subject to Apify's rules and schedule. The suggested price is not configured by this source package. As of v0.2.1, payout onboarding and PPE pricing remain pending; no buyer revenue has been independently verified.

No network calls, scraping, external agent execution, account operations or spending are performed. Input data are provided by the caller. Integrity hashes are not independent proof that a third party completed real-world work. Do not send secrets or personally identifiable information in inputs.

Recurring workflow integration

The buyer sends fresh input on each occurrence through its automation or API client; setting up a timer with unchanged sample input would generate redundant work and charges. Repeat scheduling and publication have not been configured. Each result includes a deterministic workOrder.id for customer-side reconciliation, not automatic billing deduplication.

Publish to the existing market

  1. The private Penny Engine Actor is already created in the owner’s Apify account. The code currently builds in Apify; publication is not complete.
  2. Update the configured ZIP source URL if deploying a new version; keep the Actor private until Store and payment verification is complete.
  3. Run both sample inputs before publication. The agent-QA workflow passed its first cloud test; the catalog smoke test remains to be verified.
  4. In Publishing → Monetization, choose pay per event and configure work-order-completed (start with a small, economically viable price after testing costs).
  5. Complete Apify Store's publishing checklist. Publication and payout verification are separate steps.

Local core tests use Node's built-in test runner and do not require an Apify account or any payment. To run the Apify-specific main process locally, install the Apify package (npm install) and follow Apify's Actor local-run instructions.

Penny Engine retains its existing Railway MCP gateway and x402 API service. This standalone Actor does not call or charge the existing paid Railway endpoint, preventing double-billing.

Example output and machine schema

Apify uses .actor/output_schema.json and .actor/dataset_schema.json to expose the default dataset to API clients and render the result with documented field meanings. Each work order returns a jobType, completion status, JSON and Markdown deliverables, content-integrity evidence, and stable workOrder metadata. A successful Actor run does not necessarily mean all individual checks passed; read the detailed reportJson.