Software Acquisition Scout

Research software projects worth a closer look across three public marketplaces: WordPress.org plugins, the Chrome Web Store, and the Atlassian Marketplace. Each run screens one marketplace, scores candidates with heuristic signals, flags deal risks, and reports what changed since your last scan.

Scores do not establish abandonment, availability, ownership rights, or acquisition value. They are research signals — verify before acting.

How a run works

1. Pick a marketplace and set your filters (see Input below).
2. The actor screens that marketplace, enriches each candidate, and pushes one JSON object per candidate to the run's dataset, sorted by distribution value score.
3. A RUN-SUMMARY record is written to the run's key-value store with the platform, counts (pushed / blocked by charge limit / unattempted), whether a charge limit was reached, and the input used.

Input

One run screens one marketplace. All fields are optional except platform.

Field	Default	What it does
platform	wordpress	Marketplace to screen: wordpress, chrome, or atlassian.
limit	25	Maximum candidates returned, sorted by distribution value score (1–200).
minimum_users	500	Skip assets below this install count. The scoring sweet spot is 500–20,000 installs.
minimum_rating	3.0	Skip assets rated below this (0–5). Assets with no public rating are skipped.
minimum_age_days	365	Skip assets younger than this (proven-longevity signal).
days_since_update	365	Only consider assets not updated for at least this long (abandonment signal).
maximum_competitor_count	50	Skip assets whose niche has more competing listings than this.

Example input:

{
  "platform": "wordpress",
  "minimum_users": 500,
  "minimum_rating": 3.0,
  "minimum_age_days": 365,
  "days_since_update": 365,
  "maximum_competitor_count": 50,
  "limit": 25
}

Output

Each dataset item describes one candidate with these fields:

• asset — asset name. platform — marketplace screened.
• users_installs, rating, reviews — public traction figures.
• last_release, release_frequency_trend, support_backlog — maintenance signals.
• developer_activity — version and "tested up to" information where public.
• known_monetization, competitor_pricing — how the asset and its niche make money.
• abandonment_probability — 0–1 heuristic score, not a calibrated probability. Higher means more abandonment signals; it is not proof that an asset is abandoned.
• distribution_value_score — 0–1 heuristic ranking score combining existing users, organic discovery, monetization gap, maintenance simplicity, abandonment evidence, competitor pricing, and review quality.
• estimated_technical_complexity — 1–5 maintenance estimate (lower is simpler).
• acquisition_takeover_route — how a takeover could realistically proceed on that marketplace (e.g. the wordpress.org adoption flow or direct owner transfer).
• red_flags — risk flags such as single_maintainer. Flags, not disqualifiers.
• changed_since_last_scan — installs / rating / reviews deltas, whether a new release appeared, and days between scans, computed from your cross-run history.
• source_evidence — listing URL, source URLs, competitor count, and coverage_warnings listing exactly what was unavailable on that platform.

Platform coverage

• WordPress.org — public plugin metadata; individual fields can be unavailable.
• Chrome Web Store — detail pages are JavaScript-rendered. The actor reads static HTML only: install counts are unavailable, and rating, reviews, and last-update are best-effort extractions that often miss. Expect few or zero results. Each item's source_evidence.coverage_warnings states what was missing.
• Atlassian Marketplace — public REST data. App age and support backlog are not public.

Cross-run history

Repeat scans of the same marketplace build on your history. Snapshots are kept in a named, account-scoped key-value store (scout-history-v2-…, the suffix is derived from your account and scan configuration), one immutable record per snapshot. Reads return up to 25 recent snapshots per asset; stored history may grow until removed. Only the same account with the same scan configuration shares a history store. This is what powers changed_since_last_scan. Records are retained; cleanup is a separate maintenance step and never something a scan does to your history.

Data retention

The actor does not delete these records; Apify account/platform retention applies to run datasets and default key-value store summaries. Named history is retained until removed.

Support

Questions or issues: james@jbschroeder.co

Pricing and spending limits

At up to 1 GB allocated memory, each run incurs a USD 0.50 start charge plus USD 0.10 for each scored candidate saved to the default dataset. Each additional GB increases the number of start charges. At the default 1 GB memory, zero candidates cost USD 0.50, 20 cost USD 2.50, and 25 cost USD 3.00, subject to the configured spending limit. A failed run may still incur its start charge and charges for results delivered before failure.

Set a maximum charge per run in Apify before starting. The saved pricing uses standard pay per event with platform usage included, not the optional additional platform-usage charge. Review the Pricing tab for current rates. Start with WordPress.org; Chrome coverage can produce few or no results as explained above.