# Changelog of Tech Stack Detector – Website Tech, Email Provider & Leads (`haketa/tech-stack-detector`) Actor

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

## Changelog

### 1.1.0 — 2026-11-01

- **Fixed a memory crash on large website lists.** A big fresh verification run (352 domains at concurrency 15) was terminated by an out-of-memory `SIGKILL` after ~148s, completing only 76 records. The worker pool was retaining every result in memory for the whole run even though each record is already streamed to the dataset. The pool no longer accumulates returned records, and the Actor's memory ceiling was raised to **2 GB** so large bulk runs have ample headroom. After the fix the **same 352-domain list completed 352/352 records with peak memory of just 282 MB** (down from 968 MB).
- **Verified quality on the fresh 352-record run:** core fields are richly populated — `url` 100%, `ipAddress` 100%, `sslIssuer`/`sslValidTo` 99%, `emailProvider`/`mailServers` 98%, `statusCode`/`finalUrl` 97%, `technologyNames` 95% (avg 5 technologies per site, up to 15), `title` 94%. **0 duplicates, 0 scheme-less URLs, no `[object Object]` or type bugs.** CMS (WordPress, Contentful, Sanity.io, Adobe Experience Manager, Drupal…), ecommerce platform (Shopify, Shopware…) and email provider (Google Workspace, Proofpoint, Microsoft 365, Mimecast, Cisco…) are detected accurately, and records that hit a `403`/`429` still return their DNS + SSL enrichment.
- **Refreshed the README** to the current output shape: corrected the output example and field-reference tables to match the live `inputUrl` / `url` / `finalUrl` / `technologies` / `technologiesByCategory` / `mailServers` / `error` fields exactly, clarified which signals are data-dependent, and tightened the pagination/scale and FAQ sections.
- No changes to input or output field names — fully backward compatible.
