Apify Store Opportunity Radar: Gaps, Prices & Alerts

What Apify Store Opportunity Radar does

Apify Store Opportunity Radar helps you shortlist Actor categories and monitor competitors from public Store metrics. A run reads the public Store listing, keeps aggregate numbers and derived scores, and writes a compact report. It does not copy descriptions, READMEs, images, or full listings.

The slug stays apify-store-market-intel. Scores are triage heuristics. They are not a demand forecast, a revenue estimate, or a promise of earnings.

Who it is for

Use it to:

1. Shortlist categories that deserve a closer look before you build.
2. Compare prices only when the charge unit is the same.
3. Keep your own weekly snapshot and, on a later comparable scan, read watchlist changes for Actor ids observed both times.
4. Export a short JSON or markdown brief into product planning.

Public user counts are Apify-reported users of each Actor. They are not unique people across Actors, not paying customers, and not revenue.

What a run returns

Row	What it contains
market-summary	Actors observed, reported total, pages fetched, coverage (observed-cohort or verified-complete), sum of present 30-day user stats, result-unit median with its sample size, history outcome
category-opportunity	demandScore, roomToWin, confidence, gapFlags, growth only when both weeks are verified-complete, and the top Actors' metrics
mover	Week-over-week 30-day user increases for ids in both comparable snapshots, otherwise 7-day momentum
price-change	Model changes, or unit-price changes inside the same identified unit, for ids in both comparable snapshots. Cross-unit jumps are not listed as comparable price changes
alert	Watchlist price, user, rating, or reliability changes for Actor ids in both the verified-complete baselines or both observed cohorts. dropped-from-store-listing is stored only when both weeks are verified-complete
watchlist-coverage	not-observed-this-scan when an id is outside this observed cohort. Not a removal and not an alert charge
category-deep-dive	Leaderboard and result-unit quartiles. New entrants are listed only when both weeks are verified-complete

Every run also stores REPORT, REPORT_MD, and RUN_SUMMARY in the run key-value store. You can download the dataset as JSON, CSV, or Excel.

The first saved cohort for an ISO week is kept. A later run that same week does not replace it. Price and user changes are compared only for Actor ids that appear in both scans.

Weekly history and same-week rules

The public Store list documents count as the number of items in the response and total as the size of that query. On the current API, a 1,000-item window can report count: 1000 while items.length is lower. Small pages sometimes match. total also differs between sortBy=popularity and sortBy=relevance. Item length is not treated as proof that the whole Store was seen.

Each run requests GET /v2/store with sortBy=popularity, limit=1000, and explicit offsets. There are two histories in the named store apify-store-intel-history, in the account that runs the Actor:

• Verified-complete baseline, only when every fetched page has count equal to items.length, the offset windows cover that response's total, and maxPages did not stop the crawl. Only this history can emit dropped-from-store-listing.
• Observed cohort, used whenever that rule is not met, including the current public 1,000-row pages. The record is store-intel-sample-v1. It stores the Actor ids actually returned, the sort, the offsets, and the shortfall. It does not replace a verified-complete baseline.

Weeks are ISO-8601 weeks in UTC (YYYY-Www). The key prefix is the Actor task id, or notask, plus historyKey. The sample claim also includes maxPages, so a 1-page cohort does not occupy a 30-page cohort. An ad-hoc run with no task id uses notask.

• The first snapshot of each kind in an ISO week is kept.
• A missing id on a cohort scan is not-observed-this-scan. It is not a removal, and it is not billed as an alert.
• Price, user, rating, and reliability alerts require the same Actor id in both comparable snapshots. Comparable cohorts use the same sort, page size, offsets, and maxPages.
• A missing stat stays null. It is not stored or scored as zero.
• A legacy snapshot from before this revision is not proof of a complete listing and does not raise alerts.
• On the Apify platform the named key-value store and the claim queue are opened with forceCloud, so they use the account Apify API client. If that client is missing, the run stops. It does not keep the weekly baseline on the container disk. Local runs keep using local storage. The store names are account-scoped. Task isolation stays in the key prefix {taskId}__{historyKey}.
• A complete listing claims that week by inserting one request-queue record whose uniqueKey is store-intel-claim-v2:{scope}:{week}. The queue accepts that key once. The losing run does not write a snapshot.
• The claim is not deleted, marked handled, or reclaimed. If the winning run crashes before the snapshot is stored, later runs fail instead of taking the slot. Use a new historyKey for a new baseline, or rerun only after the winning snapshot is present.
• Key-value storage holds the snapshot body. It is not used as a lock.

Schedule the same historyKey and maxPages weekly if you want the next scan to compare with this cohort. Actors that fall outside the fetched offsets are reported as not observed, not deleted.

Pricing

This Actor does not charge the platform start event itself. Apify charges apify-actor-start automatically: $0.00005 per event, one event up to and including 1 GB, then one more event per additional GB (partial gigabytes round up). Published examples: 128 MB and 1 GB are one event ($0.00005); 4 GB is four events ($0.00020). The default memory in this source is 1024 MB. A task that selects more memory, up to the 4096 MB maximum set here, scales only the start fee.

Event	Price	Charged when
Actor start	$0.00005 per GB, minimum one	Platform, at run start. Not charged again in code
snapshot	$0.02	After the market-summary row, REPORT, and REPORT_MD are stored
gap-report	$0.10	After category, mover, and price-change rows are stored, and only if includeGapReport is on
alert	$0.02	Once per stored alert row. No row means no charge
deep-dive	$0.25	Once per stored deep-dive row. At most five. An empty category is not charged

A failed run that stops before the matching artifact is stored does not charge that event. An incomplete listing can still store an honest summary and, if you asked for it, a report marked incomplete. Those stored artifacts are charged. They do not update the weekly baseline, and they do not create dropped-from-listing alerts.

Examples at 1 GB or less (one start event):

• Snapshot and opportunity report: $0.00005 + $0.02 + $0.10 = $0.12005
• That run plus 3 stored alerts and 1 stored deep-dive: $0.12005 + $0.06 + $0.25 = $0.43005

Same snapshot and report at 4 GB: start is $0.00020, so the total is $0.12020. At 2 GB the start is $0.00010 and the total is $0.12010.

Alert and deep-dive totals are the sum of the rows this run stores. A run with no alert rows pays no alert events.

How prices are compared

pricePer1kUsd is the FREE-tier list price of one result-like charge times 1,000. Result-like means PRICE_PER_DATASET_ITEM, or a pay-per-event primary event (or apify-default-dataset-item when nothing is marked primary) whose key and title identify a result or dataset item and do not mention another unit.

These units are not converted into a per-1,000 price: page, minute, report, start, gigabyte, hour, second, run, and month. Their unit price stays on unitPriceUsd and pricePer1kUsd is null. Unknown models and unidentified events are null too. The cheapest add-on is never used as the comparison price. Actor start is reported separately as startUsd and is not folded in. Rental (FLAT_PRICE_PER_MONTH) is a monthly price, not a per-1,000 price.

overpriced-leaders uses result-unit prices only, and only when the top 10 contain at least 5 positive result prices. A page-priced Actor is not ranked against a result median.

How scores are calculated

scoreVersion is 2026-11-08.

• Demand = 100 * log10(1 + knownUsers30d) / log10(1 + largestCategoryKnownUsers30d), rounded to one decimal. knownUsers30d sums 30-day user stats that are present. Missing stats are skipped, not zeroed. The same person using several Actors is counted in each of those Actors.
• Leader flags use the top 5 by 30-day users. low-rated-leaders requires at least 3 of them to have at least 3 reviews and an average rating under 4.2. unreliable-leaders requires a top-5 fail rate above 0.1 with at least 50 public runs in 30 days. stale-leaders requires at least 2 leaders whose fetched modifiedAt is older than 180 days. An unknown modified time is not stale, and a build count is not a modified time. overpriced-leaders is the result-unit rule above. fragmented-market means the top 3 hold under half of known users, with at least 5 Actors and enough known user stats. fast-growing means category known users rose more than 10% versus the previous complete snapshot.
• Room to win = demandScore * weakness * (1 - top3Share * 0.6), rounded to one decimal. weakness starts at 0.2 and adds 0.2 for each leader flag, capped at 1. Missing ratings do not increase weakness. If concentration cannot be computed, roomToWin is null.
• Confidence is low when the listing is incomplete, or when there are fewer than 5 Actors, or when fewer than 70% of Actors have a 30-day user stat. It is high only when the listing is complete, there are at least 8 Actors, and at least 90% have that stat. Otherwise it is medium when the medium thresholds are met.

Growth compares the sum of present user stats for Actors in the category this week with the same kind of sum from the previous complete snapshot. If more than 10% of either side is missing a user stat, growth is null.

How data is collected

• GET https://api.apify.com/v2/store?limit=1000&offset=... for the public listing.
• GET https://api.apify.com/v2/acts/{id} for modifiedAt on the bounded leader set only. The rest of that response is discarded.
• No other hosts, no proxies, and no redirects.
• At least 1.2 seconds between requests, including detail calls and retries. The User-Agent is ClearfeedLabs-StoreIntel/0.2 (+https://apify.com/clearfeedlabs/apify-store-market-intel).
• HTTP 429 and 5xx wait for Retry-After when it is present, otherwise for a bounded exponential backoff, never more than 60 seconds, and at most 4 attempts. If the listing still fails, the run fails. It does not treat the failure as the end of the Store.
• If the Store's reported total changes between pages (an Actor was published or removed mid-scan), the whole listing is read again once. If it changes again, the run finishes with one free run-status row that says so. Nothing is charged and history is not updated.
• The public Store API currently stops returning Actors after about offset 15,000 while still reporting the full total. An empty page after the first one ends the scan as an observed cohort (truncationReason: store-offset-limit), the same as hitting maxPages. It is never counted as a complete listing.
• The run records the pages fetched, the Actors observed, the API total, the truncation reason, and the request counts. It does not claim a fixed Store size or a fixed request count.

Input

See the input tab for every field. Sample inputs:

Snapshot only:

{ "includeGapReport": false, "maxPages": 30 }

One category report:

{ "categories": ["DEVELOPER_TOOLS"], "includeGapReport": true, "enableAlerts": false, "detailTopN": 5, "maxPages": 30 }

Watchlist from a later comparable scan:

{ "enableAlerts": true, "watchlist": ["apify/website-content-crawler"], "alertUserChangePct": 20, "historyKey": "default" }

One deep-dive:

{ "deepDiveCategories": ["DEVELOPER_TOOLS"], "deepDiveSize": 25, "includeGapReport": true }

The first watchlist run stores that week's snapshot and does not charge alert. On a later ISO week with the same historyKey and maxPages, alerts are for Actor ids present in both scans. A missing id is not-observed-this-scan, not a removal. dropped-from-store-listing is stored only when both weeks are verified-complete.

Output

Illustrative rows from the local fixture, not a live Store census:

{
  "rowType": "market-summary",
  "week": "2026-W45",
  "actorsObserved": 3,
  "reportedTotal": 3,
  "pagesFetched": 1,
  "listingComplete": true,
  "truncationReason": null,
  "users30dSum": 150,
  "users30dMissingActors": 0,
  "medianResultPricePer1kUsd": 2,
  "medianResultSampleSize": 1,
  "userCountMeaning": "Sum of Apify-reported public user counts for Actors with a present 30-day stat. Not unique people, not paying customers, and not revenue. Missing stats are excluded, not counted as zero.",
  "nonAffiliation": "Not affiliated with, endorsed by, or sponsored by Apify.",
  "source": "illustrative fixture, not a live Store census",
  "fixture": true
}

{
  "rowType": "category-opportunity",
  "week": "2026-W45",
  "category": "DEVELOPER_TOOLS",
  "demandScore": 100,
  "roomToWin": 8,
  "confidence": "low",
  "gapFlags": [],
  "users30d": 150,
  "medianResultPricePer1kUsd": null,
  "resultPriceSampleSize": 1,
  "listingComplete": true
}

A page-priced Actor in that output has priceUnit page, unitPriceUsd set, and pricePer1kUsd null. It is not flagged overpriced against result prices.

Other Store research Actors

Prices below were inspected on November 8, 2026 and can change. User counts are not proof of buyer demand.

Actor	What to expect	Price read that day
zinin/apify-market-gap-finder	Bounded niche cohorts, ranked opportunities, comparator evidence, confidence, data gaps, and a manual-validation step	$0.10 per delivered report, plus $0.005 per GB start with a minimum of one start. The listing says invalid, insufficient-evidence, failed, and replay outcomes are not charged as that report
conversational_kermis/apify-market-pricing-intelligence	Pricing recommendations, tier curves, competitor and rising-star notes, gaps, and scenarios	$0.01 per dataset result and $0.01 per strategic summary

This Actor's distinct behavior is the account-owned weekly baseline, watchlist rows, and same-unit price handling described above. It does not claim a better model, unique coverage, or measured performance against those Actors.

Disclaimer

Not affiliated with, endorsed by, or sponsored by Apify. Apify is a trademark of its owner. Metrics come from Apify's public Store API and can be delayed, rounded, incomplete, or changed.

Scores and flags are automated heuristics. Check the Actor's Store page before you act on a row. This is not financial, legal, or business advice. Do not use the output to copy another developer's Actor, content, or branding.

Questions and corrections belong on this Actor's Issues tab.