# Changelog of Clutch.co Scraper — B2B Agencies, Ratings & Reviews (`blackfalcondata/clutch-scraper`) Actor

- **URL**: https://apify.com/blackfalcondata/clutch-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/blackfalcondata/clutch-scraper.md

## Changelog

### 2026-10-17

#### Fixed

- Runs no longer fail with "Run failed. Please try again later." when clutch.co shows its "checking your browser" page in a language other than English. The scraper now waits for that page to pass whatever language it is in.

### 2026-10-16

#### Fixed

- Every run now reports how many companies it delivered as soon as it finishes, so a successful run is no longer shown as empty in the Actor's health checks while its count is still being settled.

### 2026-10-12

#### Fixed

- A run with a **maximum cost per run** now returns the companies that cost covers and ends as succeeded, instead of being stopped by the platform part-way. A company the limit held back is still reported as new or changed on the next incremental run.
- Runs are steadier on large scans: memory use now stays within the run's memory setting.
- Default run memory raised to 2 GB so the run starts reliably. Runs at the 512 MB default set on 2026-09-30 could fail right after starting.
- If the scraper's browser cannot start, it is now restarted once, and a run where it still cannot start says so, including the memory to use, instead of reporting that the service is temporarily unavailable.
- The input form again comes pre-filled with an example listing URL, so starting a run with the defaults returns companies instead of nothing.
- A small **maximum cost per run** (for example $0.05) now returns the companies it pays for. Such runs could end with no results and the message that the cost left nothing after the start fee.

### 2026-09-30

#### Changed

- Default run memory lowered from 4 GB to 512 MB, which covers the full run. Default runs are cheaper to start and use less platform compute.

All notable changes to this Actor are documented here.

### 0.1 — no more silent empty runs

- A run started without a Clutch URL or a search term now says so, instead of reporting "no matching results".
- A directory page that loads without any companies on it is now fetched again
  before it is trusted. If it still shows none, the run reports that the source
  could not be read instead of finishing successfully with 0 results.

### 0.1 — search and result-limit contract repair

- Added keyword search through Clutch's current provider-search surface.
- Restored `startUrls` as a hidden backwards-compatible alias for `searchUrls`.
- Added a global `maxResults` control across all URLs while preserving the
  existing `maxItemsPerUrl` control for each individual URL.
- Added parsing and pagination for the compact provider cards returned by
  Clutch search.

### 0.1 — initial release

- Scrape B2B service providers from any Clutch category or search URL:
  company name, verified badge, star rating, review count, minimum project
  size, hourly-rate band, employee-count band, office location, service-focus
  breakdown, short description, project count, website and profile link.
- Opt-in **Include company details** adds each company's bio, social links,
  office locations, technology stack and industry focus.
- Opt-in **Include reviews** adds each company's published client reviews.
- **Exclude empty fields** option to omit null/empty fields from each record.
- Optional run-completion notifications via Telegram, Discord, Slack or a
  custom webhook.
- Reliable, repeatable runs over the same search across time.