# Changelog of BuiltWith Alternative — Tech Stack Leads & Lookup (60M) (`apivault_labs/tech-stack-lookup`) Actor

- **URL**: https://apify.com/apivault_labs/tech-stack-lookup/changelog.md
- **Full Actor documentation**: https://apify.com/apivault_labs/tech-stack-lookup.md

## Reliability repair — 2026-10-05

- Structured SUMMARY/ERRORS for exports, counts, seed and monitor lookups. All-source failure is FAILED; healthy empty matches remain successful.
- Preserve monitor baseline and suppress change events on incomplete lookup. Suppress lookalike suggestions from incomplete seeds.
- SQL statement_timeout=30s, lock_timeout=5s, connect_timeout=10s.
- Invalid input writes machine-readable error records without fabricated results.

## Changelog

### \[0.14.3] — 2026-09-18

#### Added

- AI/MCP-oriented output presets: essential, contacts, technology stack, sales
  opportunity and full export. Essential is now the compact default; an exact
  custom column list still overrides the preset.

#### Changed

- Reworded the input form around five explicit user workflows and labeled every
  workflow-specific field, reducing ambiguity for both humans and agents.
- Added root-level MCP routing guidance, recommended interactive limits and
  explicit rules for choosing technology presets versus custom values.
- Invalid workflow inputs now fail with a precise actionable message instead of
  looking like a successful run with an unexplained empty dataset.
- Expanded Store metadata for technology-lead, BuiltWith-alternative and
  find-websites-by-technology discovery queries.

### \[0.14.2] — 2026-09-18

#### Fixed

- Query schema variants safely: fields absent from an older physical table are
  returned as null instead of excluding the entire data segment.
- Apply filters only to columns that physically exist in each segment, so a
  heterogeneous legacy table cannot fail a complete export.
- Scale per-segment query pages to the requested result limit and increase the
  bounded read concurrency, reducing first-result latency for small exports.
- Preserve the encrypted production `NEON_SHARDS` registry during source-only
  redeploys and refuse to create a new version from an incomplete local copy.
- Make the release audit follow the actual latest build instead of a hard-coded
  historical build number.

### \[0.14] — 2026-09-18

#### Fixed

- Query every configured data segment, including segments whose physical table
  name differs from their public platform group.
- Ignore database-pool provisioning metadata so physical databases are not
  queried twice.

### \[0.13] — 2026-09-10

#### Added

- Six ready-made sales opportunity presets for stack-gap prospecting.
- Opportunity type, missing technologies, opportunity score, signal strength
  and a campaign-ready suggested pitch on matching leads.
- Buying-signal classification for monitored changes: technology adoption,
  removal, general stack change and same-category migration.
- Migration From/To fields and a dedicated Stack Gap Opportunities view.
- Agent-first mode routing, MCP configuration, typed dataset fields, a safe
  example input and explicit limited-permission execution.
- External MCP discovery benchmark and a public 10-result opportunity demo Task
  linked directly from the first-screen call to action.

#### Fixed

- Opportunity presets now search only their relevant stack layer, preventing
  matches such as Shopify Pay from being presented as Shopify-store prospects.
- Required and missing technologies in an opportunity preset can use different
  stack layers, so Shopify is checked in ecommerce while Klaviyo is excluded
  from marketing automation.

### \[0.12] — 2026-09-10

#### Added

- **Lookalike Leads** mode: provide seed customer domains and find companies
  sharing their strongest known technology signature.
- Public technology catalog plus a popular-technology preset and alias
  normalization for common shorthand such as WP, Woo, GCP and Hub Spot.
- Derived lead intelligence: matched technologies/categories, technology count,
  fit and contactability scores, lead tier, explanation, data age and similarity.
- Dedicated Lead Intelligence dataset view.
- Persistent **Monitor technology changes** mode. The first run saves a baseline;
  later runs emit and charge only added/removed-technology change rows.
- Optional live verification with private-network protections for up to 100 known domains, including
  HTTP status, current technology signals, categories, confidence, evidence and
  technologies observed live but absent from the stored profile.

### \[0.11] — 2026-09-10

#### Added

- A second **Check my domains** workflow for domain-to-technology lookup and
  lead enrichment, alongside the existing reverse technology search.
- Exact, normalized domain matching for URLs and hostnames; unknown domains are
  skipped and do not create billable results.
- Faster exact-domain checks without changing reverse-search behavior.

#### Changed

- Repositioned the listing around high-intent **BuiltWith alternative**,
  **Wappalyzer alternative**, **tech stack lookup** and **technographics** terms.
- Reduced the default first run from 1,000 to 50 rows so users can evaluate
  data quality with a low-cost sample.
- Added Developer Tools as a Store category and introduced $5/1K launch pricing.

### \[0.10] — 2026-09-09

#### Added

- Freshness filters for websites first indexed, last indexed or last confirmed
  within a configurable number of days.
- First Indexed and Last Indexed as global sort choices.
- A dedicated Freshness dataset view.
- Five reusable Task presets for recent Shopify, Klaviyo, HubSpot, Stripe and
  WordPress lead searches, mostly using a rolling seven-day new-lead window.
  Tasks are not automatically scheduled.

#### Fixed

- Future-dated source anomalies are excluded from freshness filters and treated
  as missing values during date sorting.

### \[0.9] — 2026-09-09

#### Fixed

- Root Domain is displayed as readable text instead of an invalid URL error in
  every dataset view.
- Results are now merged and sorted globally across all data segments instead
  of filling the result limit from the first matching segment.
- Numeric fields such as Employees, Sales Revenue and Overall Score use numeric
  ordering rather than text ordering.
- Duplicate removal happens globally before offset pagination; it is enabled by
  default and normalizes scheme, `www.` and trailing slashes.
- Count-only preview and invalid input no longer create per-result dataset charges.

#### Changed

- Unsorted queries return a fast, balanced round-robin mix instead of favoring
  the first data segment. Explicit sorting is merged globally and uses Root
  Domain as a stable tie-breaker.
- Count-only results are stored in the non-result-billed `COUNT_SUMMARY` record and
  clearly report raw segment rows as an upper bound, not unique websites.
- `_platform` is labeled **Data segment** to avoid presenting it as the site's
  detected technology.

### \[0.8] — 2026-09-07

#### Changed

- **Per-run limit raised 100K → 250K rows.** There is no total cap — use Offset
  to page through unlimited results across runs without paying twice. When a run
  returns the maximum, the status message now shows the exact Offset to use next.

### \[0.7] — 2026-09-07

#### Changed

- Replaced competitor-comparison wording with a distinct product category:
  **technology lead finder** and **B2B lead lists by technology**.
- Kept high-intent search language focused on finding websites that use a
  specific technology.

### \[0.6] — 2026-09-07

#### Added

- New original README hero illustrating the search-to-lead-list workflow.
- New matching square Actor icon, optimized for a clear small-size silhouette.

### \[0.5] — 2026-09-07

#### Changed

- Redesigned the input form into four guided sections: technology targeting,
  lead filters, lead prioritization and export settings.
- Added clear emoji labels, examples and recommended choices so first-time users
  can build a technology lead list without reading the full documentation.
- Moved output-column selection to the final export step for a more natural flow.

### \[0.4] — 2026-09-07

#### Changed

- Expanded the Store guide with quick starts, complete input documentation,
  campaign playbooks, API examples, export guidance, data-quality notes and FAQ.
- Improved search coverage for technology lead lists, websites using Shopify,
  Klaviyo and HubSpot, and technology-based prospecting.
- Simplified public wording so the documentation explains product behavior and
  limitations without exposing implementation details.

### \[0.3] — 2026-09-07

#### Changed

- Repositioned the Store listing around the buyer intent: **find websites using a
  technology**, **technology lead lists**, and **B2B technology lead finder**.
- Improved the title, short description, input labels, use cases and FAQ for
  Apify Store and external search discovery.
- Clarified the reverse technology lookup workflow and data-freshness expectations.

### \[0.2] — 2026-09-07

#### Added

- **Multiple technologies at once** with a match mode — **Any** (OR, e.g. Stripe
  or Braintree) or **All** (AND, e.g. Shopify + Klaviyo).
- **Exclude technologies** — remove sites using a given tool (e.g. Klaviyo users
  that do NOT use Yotpo).
- **Industry / vertical filter** — narrow results to an industry (Fashion,
  Technology, Health…).

### \[0.1] — 2026-09-07

#### Added

- Initial release. Find websites using a given technology across **60M+ sites**.
- Search by **Klaviyo, Stripe, Shopify Plus, HubSpot, Cloudflare, PayPal** and
  any other tool in the ecommerce / CMS / CRM / marketing / payments / hosting /
  AI stack — matched across all layers by default or narrowed to one.
- 59 enriched fields per site: emails, phones, full tech stack, firmographics,
  social profiles, rankings and geo.
- Filters: country (multi-select), only-with-email, only-with-phone.
- Sort by key metrics, deduplicate by domain, count-only preview, offset paging.
- Output column selection — leave empty for all 59 fields.
- Pay per result; 1 site = 1 row. Designed for reliable large exports.