Purchase Offer Validator

Check whether a structured offer meets the purchasing requirements your agent already holds. Process up to 100 comparisons in a batch and receive structured results suitable for a workflow branch, review queue or audit record.

This tool checks exact product identifiers and variants, item condition, quantity, currency, destination reference, reported stock, order-total arithmetic, budget, quote freshness, delivery date and return-window length. An optional previous offer exposes changed fields and price differences.

Try it

Run {"demo": true, "checks": []} for three fictional examples: a match, a refurbished-item mismatch and unknown shipping. These samples trigger no purchase-checked event fee. Any platform charges shown in Console remain applicable. The demo never checks your own purchases.

For your own offers, send {"demo": false, "checks": [PAYLOAD]}. Each PAYLOAD contains requirements, offer, and optionally previous_offer. The complete JSON Schema is included as PURCHASE_SCHEMA.json. The demo output is a result, not a reusable current quote; supply actual offer facts from your authorized source.

Money must be decimal strings. Cost components are whole-order amounts: subtotal + shipping + tax + fees - discount = total. Quantities must already be normalized to the same sellable unit; this Actor does not infer pack sizes or minimum orders. Use one identifier namespace and exact variant attributes. Use opaque destination references, not addresses. Unsupported facts must be null, not invented values. Supplied timestamps must have a timezone.

Results

One dataset row per completed comparison, in input order:

• matches_supplied_requirements: supplied facts satisfy the requirements.
• mismatch: at least one supplied fact conflicts with a requirement.
• needs_evidence: necessary information is missing or stale.

Rows include input_index, issues, changes, computed_total, reported_total, input_sha256, and payment_authorized: false. OUTPUT in the run's key-value store summarizes delivered comparisons and spending-limit stops.

Charging

Price: $0.01 per completed comparison (purchase-checked event). One completed non-demo comparison triggers one event, including mismatch and needs-evidence results. Invalid batches are rejected before any comparison event. There is no event charge for fictional demos.

Set a maximum run charge in Console or the API. The Actor stops when its SDK reports the event budget is exhausted. Results already delivered remain in the dataset. Check OUTPUT and the dataset before retrying. Restarting an already-started run is refused to reduce duplicate processing; starting a separate run is a new request and can incur a new charge. Interrupted storage/payment operations are not guaranteed atomic and must be reconciled before retries.

Scope and privacy

The Actor does not scrape, retrieve prices, verify merchant claims, parse documents, place orders, authorize payment, establish product compatibility, or guarantee delivery/returns. A matching result is not permission to spend. It needs no merchant credentials, browsing proxy or LLM key.

Apify stores run inputs and result datasets under its platform settings; this hosted version is not local-only. Do not submit credentials, payment-card details, customer identities, street addresses or confidential source URLs. Result rows can contain supplied offer IDs, source references and changed values. Keep run datasets private and control retention through your Apify account. We do not add a separate analytics or data-resale pipeline.

Maximum input: 1 MB and 100 comparisons. Execution is batch-based, not a low-latency payment gate. Caller-submitted strings and URLs are treated as data, not fetched or executed.

For agent integration, call this Actor through the Apify API or an Apify MCP tool configured by your operator. Use the returned status and issues, while keeping purchase authorization in your existing workflow. Report reproducible issues with synthetic/redacted examples through the listing's Issues tab.