Capterra Reviews Scraper

Capterra reviews scraper that extracts complete, structured review data — overall and sub-ratings, pros, cons, reviewer role, industry, company size, verified status, and vendor responses — from any Capterra product page, plus full product listings from any Capterra category. You are charged per successfully extracted review. Blocked or empty results are never billed.

Last updated: July 2026 · actively maintained.

Why this one

• No empty charges — ever. A result is billed only after it has been extracted, validated, and written to your dataset. Runs that hit anti-bot walls or return nothing cost you nothing. That is a guarantee, not a goal.
• Every field, every row. Most Capterra scrapers return a title and a star count. This one returns the full review object: five sub-ratings, pros/cons/overall comment, reasons for choosing and switching, alternatives considered, reviewer role, industry, company size, usage duration, LinkedIn-verified flag, and vendor responses.
• Built for survival, not best-case runs. Reviews are read from Capterra's embedded structured data (more stable than HTML scraping), with an independent DOM fallback, a real-browser engine that clears Capterra's anti-bot checks, rotating residential sessions, block detection, and exponential backoff. When Capterra changes cosmetics — or defenses — this actor keeps working.

What it does

Mode A — Product reviews. Give it Capterra product URLs; it extracts every review (or up to your limit) with all fields below.

Mode B — Category discovery. Give it a category URL (e.g. https://www.capterra.com/collaboration-software/) or a search term (matched against Capterra's 900+ category directory); it enumerates products with name, rating, and review count — and can optionally fan out and scrape each product's reviews too.

Input examples

Scrape reviews for one or more products:

{
    "mode": "product_reviews",
    "productUrls": ["https://www.capterra.com/p/135003/Slack/reviews/"],
    "maxReviewsPerProduct": 200,
    "sortReviewsBy": "most_recent"
}

Discover all products in a category:

{
    "mode": "category_discovery",
    "categoryUrl": "https://www.capterra.com/collaboration-software/",
    "maxProducts": 200,
    "scrapeReviewsForDiscovered": false
}

Output schema (one item per review)

Field	Type	Notes
source	string	Always "capterra"
productId / productName / productSlug / productUrl	string	Product identity
productCategory	string|null	From the page's breadcrumb
productAvgRating / productReviewCount	number|null	Product-level aggregates
reviewId	string	Stable Capterra review id
reviewUrl	null	Capterra exposes no per-review permalink — null, never fabricated
title	string|null	Review headline
overallRating	number	1–5
subRatings.easeOfUse	number|null	1–5
subRatings.customerService	number|null	1–5
subRatings.features	number|null	1–5
subRatings.valueForMoney	number|null	1–5
subRatings.likelihoodToRecommend	number|null	0–10
pros / cons / overallComment	string|null	Full review text
adviceToOthers	string|null
reasonsForChoosing / reasonsForSwitching	string|null	Purchase context
alternativesConsidered / switchedFrom	string[]	Product names
incentivized	string|null	e.g. "NominalGift" if Capterra disclosed an incentive
reviewerName	string|null	As displayed publicly by Capterra
reviewerRole / reviewerIndustry	string|null
companySize	string|null	Normalized buckets: self-employed, 1-10, … 10001+
usageDuration	string|null	e.g. "1-2 years"
verified	boolean|null	Capterra's reviewer-validation flag
verifiedLinkedIn	boolean|null
anonymous	boolean|null
sourceOfReview	string|null	e.g. "Capterra"
publishedAt	string|null	ISO-8601 date
helpfulCount	null	Not exposed by Capterra — null, never fabricated
vendorResponse	string|null	Vendor's reply text if present
scrapedAt	string	ISO-8601 timestamp

Category discovery items have "type": "product" with productId, productName, productSlug, productUrl, productAvgRating, productReviewCount, and description.

Missing data is always null — fields are never guessed or fabricated.

Pricing, in plain words

You pay a flat rate per successfully extracted review (and a smaller rate per discovered product in category mode). Before anything is billed, the row must pass a validation gate: it must contain an overall rating and real review text. Requests that get blocked, return empty pages, or fail validation are retried or skipped — they are never written to your dataset and never charged. There is no rental fee and no charge for starting a run that finds nothing.

Sorting and filtering

• sortReviewsBy: most_helpful (default) is Capterra's own order and is exact.
• most_recent, highest, lowest sort the reviews the run collected. Capterra serves reviews in most-helpful order, so with a maxReviewsPerProduct cap the run fetches that many reviews first, then sorts them. To get a complete date-ordered set, leave maxReviewsPerProduct at 0 (all reviews).
• minRating / maxRating / postedWithinDays filter rows before they are saved (and before they are billed).

Limitations

• Scope is capterra.com (the primary/US site). International Capterra domains are not covered in v1.
• Reviews are fetched in Capterra's most-helpful order; see the sorting note above.
• Per-review permalinks and helpful-vote counts are not publicly exposed by Capterra, so those fields are null rather than invented.
• Very large products (20,000+ reviews) work but take proportionally longer; use maxReviewsPerProduct if you don't need the full archive.

Compliance

This actor accesses only publicly visible review pages — no login, no paywall or authentication bypass, and no personal data beyond what Capterra itself displays publicly (reviewer display name, role, company size). It does not attempt to de-anonymize reviewers who chose anonymity. Please use reasonable volumes and respect Capterra's terms of service and applicable data-protection laws; you are responsible for how you use the extracted data.