FreshChoice NZ Scraper | $4.00/1K
Pricing
from $3.20 / 1,000 products
FreshChoice NZ Scraper | $4.00/1K
FreshChoice New Zealand (freshchoice.co.nz) product scraper: name, pack size, price in NZD, was price, unit price, special flag, tags, image and product URL, for any of about 80 FreshChoice stores. By keyword or category, with continuation. Pay per result, no login.
Pricing
from $3.20 / 1,000 products
Rating
0.0
(0)
Developer
architek.earth
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
2 days ago
Last modified
Categories
Features
Share
What's new, 2026-11-16
Bounded examples, explicit costs and integration recipes. NZ pricing is unchanged; no launch price cut applies.
Recorded run: wWoXLDbuWLjrBsGIv on 2026-11-15 returned 3 dataset rows; 3 primary events were charged. Status: SUCCEEDED. Dataset rows can include free status records.
Architekture, an independent data studio. architek.earth.
FreshChoice NZ returns selected-store prices, specials and unit prices. Use the daily example to collect repeatable snapshots; use incremental mode only where the input exposes it.
Tested this week
| Check | Our recorded sample | Field context, 2026-11-16 |
|---|---|---|
| Evidence window | 2026-11-15; within the last seven days | Rolling 30-day public counters across 1 title-bucket members |
| Outcome | SUCCEEDED; 3 dataset rows; 3 charged primary events | 0.00% non-success over 1 runs |
| Cost | $0.001410 platform cost in the retained run | Not a competitor execution or a matched-input benchmark |
Field source: public Store API. Non-success includes failures, aborts and timeouts. The field total is not a unique-user count. These different workloads do not establish that our actor is more reliable.
What, data and cost
One primary product event costs $0.004; 1,000 cost $4.00. Ten primary events plus one collection-delivered cost $0.19. Optional events below are additional; status rows are free.
$5 free credit ≈ 1,212 products at the primary launch rate after one initial collection-delivered charge. This estimate excludes optional event charges and additional collection starts. Credit pays for billed work; it is separate from the free preview.
Set preview: true. One bounded preview returns up to 10 data rows when the source has results. Preview rows are not charged. A preview never silently switches to paid collection.
Event prices
| Event | USD per event | When charged |
|---|---|---|
product | 0.004 | Delivered primary result |
collection-delivered | 0.15 | Once per successful collection with delivered rows |
Free and Bronze use the full primary rate; Silver receives 10% off, and Gold, Platinum and Diamond receive 20% off. Secondary event prices do not change. Your run's effective price is shown before collection.
NZ pricing is unchanged; no launch price cut applies.
Worth knowing up front
- Prices are store-scoped and come from the store's own online shop. The shops only answer New Zealand connections, so collection goes through a New Zealand residential address; that cost is included in the price.
- Search terms and category pages of the online shop are supported, 48 products per page.
- Not returned: barcodes, stock levels, nutrition or ingredients.
availabilityis always"unknown". - A run stopped by its time bound or spending limit keeps what it delivered and saves
CHECKPOINT; a run that stops early writes one free status row ("recordType": "status") that says why.
Input recipes
1. Preview: Milk prices
Open the configured example task
{"searchTerms": ["milk"],"maxItemsPerQuery": 1,"storeId": "queenstown","preview": true,"continuePreviousRun": false}
2. Daily: Milk prices at the selected store
Open the configured example task
{"searchTerms": ["milk"],"maxItemsPerQuery": 1,"storeId": "queenstown","preview": false,"continuePreviousRun": false}
3. One-off: Bread prices
Open the configured example task
{"searchTerms": ["bread"],"maxItemsPerQuery": 1,"storeId": "queenstown","preview": false,"continuePreviousRun": false}
4. Daily: Butter prices at the selected store
Open the configured example task
{"searchTerms": ["butter"],"maxItemsPerQuery": 1,"storeId": "queenstown","preview": false,"continuePreviousRun": false}
Recurring collection
Copy the example Daily: Milk prices at the selected store into your account. Schedule it in UTC with:
0 9 * * *
The worked ten-result run above repeated daily costs about $5.70 per 30 days before optional events and discounts. These tiny recipes demonstrate the input; increase their row limits deliberately. Keep one watch identifier per consistent filter set where incremental mode is available.
Output and limits
Read the live input and output definitions. Keep source identifiers, capture time and any status fields when joining snapshots. The data fields remain additive; this listing change does not promise new source data.
MCP config for Claude and Cursor
Claude: add a custom connector with URL https://mcp.apify.com?tools=architekture/freshchoice-nz-scraper and complete its sign-in. For Cursor, add this entry to your MCP settings and set your own APIFY_TOKEN:
{"mcpServers": {"freshchoice-nz-scraper": {"type": "http","url": "https://mcp.apify.com?tools=architekture/freshchoice-nz-scraper","headers": {"Authorization": "Bearer ${env:APIFY_TOKEN}"}}}}
n8n and Make
Python example and n8n workflow. In n8n, select a Header Auth credential with header Authorization and your bearer token. The workflow is inactive until you choose to run or schedule it.
In Make, add an HTTP request: POST https://api.apify.com/v2/acts/architekture~freshchoice-nz-scraper/run-sync-get-dataset-items?timeout=120&memory=2048&maxTotalChargeUsd=0.1780000. Set Content-Type: application/json, configure your bearer-token credential, and paste one recipe as the JSON body. Parse the returned array and route status rows separately. Add a scheduler only after inspecting the first result. Do not automatically retry an ambiguous POST; first check whether a run was created.
FAQ
The linked public field issues are context for these checks, not claims that the same defects exist in this actor. Grocery examples also use adjacent supermarket issues; tennis uses the adjacent tennis-data field.
Why can a green run contain no usable rows?
A completed run can still report an empty or blocked source. Read the status rows and run summary before treating it as a successful observation; a missing result is not a measured zero. Observed field issue.
Why did I receive fewer rows than the limit?
The limit is a ceiling, not a guaranteed count. The source may have fewer public matches, filters may remove records, or the run may stop at a time or spending cap. Start with maxItemsPerQuery. Observed field issue.
Can this export the whole source?
The five examples are deliberately bounded. A title or catalogue estimate does not prove full coverage. Use only the collection and continuation modes documented in the input, and retain any partial-run status. Observed field issue.
Which direct identifiers or URLs are accepted?
The input exposes categoryUrls. Use their documented formats; a source URL is not automatically a valid value for a keyword field. Observed field issue.
Why is an output field absent?
The source must publish a value for the actor to return it. Do not infer a missing price, salary, rating, ID or language. Check the output definition and preserve null or absent values in downstream storage. Observed field issue.
How do I avoid counting a repeated snapshot twice?
Treat each ordinary run as a snapshot. Deduplicate downstream using the source identity and capture time; this listing does not claim cross-run deduplication where the input has no incremental mode. Observed field issue.
What should I check after a source or proxy error?
Inspect the run message and the smallest configured example. Do not substitute login cookies or unrelated proxy credentials. A source wall is a coverage limit; preserve partial output and avoid treating failed checks as changes. Observed field issue.
How do I bound the bill?
Set a run charge cap, for example $0.178 for these tiny inputs, and keep row limits small. Ten primary results plus one collection/root-post charge cost $0.19 at the listed rate. Optional events and repeated checks add cost. Observed field issue.
Does the preview switch to paid collection?
Set preview: true. One bounded preview returns up to 10 data rows when the source has results. Preview rows are not charged. A preview never silently switches to paid collection. Observed field issue.
What evidence helps diagnose a changed result?
Provide the run ID, intended country/store/date or filter, and one expected public source URL. Include the status message and missing field name. Never post credentials. The feedback input records a note and does not change output. Observed field issue.
For AI agents
Call architekture/freshchoice-nz-scraper with a bounded recipe and an explicit spending cap. Read the input schema first, request only supported fields, and inspect status rows before summarizing results. Preserve source IDs and capture times. Never interpret absent source data as a confirmed change. Use the same watchId for repeat observations only when this actor exposes incremental mode.
Run health, reviewed 2026-11-16: last retained run wWoXLDbuWLjrBsGIv, SUCCEEDED, 3 dataset rows, observed 2026-11-15. No new run was started for this listing release.
More from architek.earth
Support and legal
Use this actor's issue tab for product questions. Public-source access does not remove your responsibility for permitted use, copyright and privacy. This is an independent product, not an official source API. architek.earth.
Changelog
- 2026-11-16: Bounded example tasks, dated health comparison, explicit worked costs, MCP and automation examples, and compatible input-name changes where applicable. NZ pricing is unchanged; no launch price cut applies.
Earlier releases
-
2026-11-16 : Support approved primary-event Store discount tiers; retain exact event checks and existing preview limits.
-
2026-11-15 : 0.1.2 first public release (runtime identical to canary build 0.1.1; README only).