Capterra Reviews Scraper

Extract every review from any Capterra product page — ratings, pros and cons, reviewer job title and company size — as clean, ready-to-use JSON, CSV, or Excel.

What this Actor does

Give it a Capterra product URL (or just its product ID/name), and it returns one structured record per review: overall rating, the five sub-ratings Capterra collects (ease of use, customer service, features, value for money, likelihood to recommend), the written pros/cons/comments, and reviewer context (role, industry, company size) when available. Reviews are deduplicated, dates are normalized to ISO 8601, and every field is either a real value or null — never undefined, never inconsistent shapes between rows.

If a product doesn't exist or Capterra changes its page layout mid-run, you get a readable error row explaining what happened instead of a silent empty dataset or a crashed run.

Use cases

• Competitive intelligence for SaaS teams — track what customers say about competitors' ease of use, support quality, and pricing perception, updated on a schedule.
• Market research — compare category-wide satisfaction (e.g. "CRM software") across dozens of vendors using consistent, comparable fields.
• Sentiment and feature-request analysis — feed the pros, cons, and comments fields into an LLM or NLP pipeline to mine recurring complaints and feature requests at scale.
• Lead generation — the reviewerCompanySize and reviewerIndustry fields help identify which segments are actively evaluating tools like yours.

Input

Field	Type	Description
startUrls	array	Full Capterra review-page URLs, e.g. https://www.capterra.com/p/135003/Slack/reviews/.
productSlugs	array	Alternative to startUrls: "{productId}/{productName}" pairs, e.g. "135003/Slack".
maxReviewsPerProduct	integer	Cap per product. 0 = no limit. Default 200.
sortBy	enum	most_recent (default), most_helpful, highest_rating, lowest_rating.
minRating / maxRating	integer	Keep only reviews within this overall-rating range (1-5).
includeReviewerDetails	boolean	Set false to omit reviewer name/role/industry/company size. Default true.
proxyConfiguration	object	Apify Proxy config. Residential proxy is required — verified: datacenter IPs are blocked by Capterra on every attempt. See Troubleshooting.

You need either startUrls or productSlugs (or both) — Capterra's robots.txt disallows crawling /search, so there's no way for the Actor to resolve a plain company name to a product ID on its own. Copy the ID and name straight from a Capterra URL you already have.

Example input:

{
    "startUrls": [{ "url": "https://www.capterra.com/p/135003/Slack/reviews/" }],
    "maxReviewsPerProduct": 500,
    "sortBy": "most_recent",
    "minRating": 1,
    "includeReviewerDetails": true,
    "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Output

One dataset item per review:

Field	Type	Description
productName	string | null	Product display name.
productSlug	string	"{productId}/{productName}" from the source URL.
productUrl	string	Canonical reviews page URL.
reviewId	string | null	Capterra's own review ID when the page exposes one (the normal case); otherwise a stable ID this Actor derives from the review's content, so dedup across pages/runs still works either way.
reviewUrl	string | null	Always null - confirmed live that Capterra does not expose a per-review anchor or link anywhere on the page. Kept in the schema for shape stability rather than removed outright.
title	string | null	Review headline.
overallRating	number | null	1-5.
ratingBreakdown	object | null	{ easeOfUse, customerService, features, valueForMoney, likelihoodToRecommend }, each 1-5 (likelihoodToRecommend is 0-10, matching Capterra). Reliably populated in the normal case; null only if this Actor had to fall back to Capterra's JSON-LD data, which doesn't carry sub-ratings.
pros / cons / comments	string | null	Free-text review content (comments is the reviewer's general write-up, separate from the pros/cons fields).
reviewerName / reviewerRole / reviewerIndustry / reviewerCompanySize / reviewerUsedFor	string | null	Reviewer context (null when includeReviewerDetails is false). reviewerCompanySize is only available in the normal case - Capterra's rendered review cards don't show it at all, so this Actor's DOM fallback path can't recover it if it's ever needed.
isVerified	boolean | null	Whether Capterra marked the reviewer as verified (their own isValidated flag - covers LinkedIn-, business-email-, and manual-proof-of-link-verified reviewers alike). Only available in the normal case; null if a fallback path had to be used.
publishedAt	string | null	ISO 8601.
helpfulCount	number | null	Always null - confirmed live that Capterra does not show "helpful" vote counts anywhere on this page. Kept in the schema for shape stability rather than removed outright.
scrapedAt	string	ISO 8601 timestamp of the scrape.
errorType	string | null	null for a normal review; otherwise PRODUCT_NOT_FOUND, BLOCKED, PAGE_STRUCTURE_CHANGED, NO_REVIEWS_FOUND, or FETCH_FAILED.
errorMessage	string | null	Human-readable explanation when errorType is set.

Example item (real output from a live run):

{
    "productName": "Slack",
    "productSlug": "135003/Slack",
    "productUrl": "https://www.capterra.com/p/135003/Slack/reviews/",
    "reviewId": "Capterra___7203282",
    "reviewUrl": null,
    "title": "Slack is very slack",
    "overallRating": 3,
    "ratingBreakdown": {
        "easeOfUse": 3,
        "customerService": 2,
        "features": 2,
        "valueForMoney": 2,
        "likelihoodToRecommend": 1
    },
    "pros": "It showed all available team members and others in the organization to reach out and connect with to collaborate.",
    "cons": "Not easy to navigate and cumbersome. Too noisy for my taste in terms of what message was being conveyed.",
    "comments": "I personally did not like the platform. It was overwhelming for me and difficult to navigate to get the needed information.",
    "reviewerName": "Ruth G.",
    "reviewerRole": "Project Manager",
    "reviewerIndustry": "Insurance",
    "reviewerCompanySize": "51-200 employees",
    "reviewerUsedFor": "6-12 months",
    "isVerified": true,
    "publishedAt": "2026-08-22T00:00:00.000Z",
    "helpfulCount": null,
    "scrapedAt": "2026-09-01T03:17:12.169Z",
    "errorType": null,
    "errorMessage": null
}

An error row (e.g. product not found) looks the same shape, with the review fields null and errorType/errorMessage filled in — filter on errorType == null to get only real reviews.

Pricing

This Actor uses Apify's pay-per-event pricing — no monthly subscription, no platform or compute fee on top:

Event	Price
Run started	$0.005 per run
Review scraped	$0.003 per review ($3 per 1,000 reviews)

Filtered-out, deduplicated, or over-your-limit reviews are never charged. A run that gets blocked and returns nothing costs you half a cent.

Example: pulling every review for 5 products at ~400 reviews each = 2,000 reviews = $6.01.

Troubleshooting

"BLOCKED" error rows / run finishes with zero reviews. Capterra serves a 403 anti-bot challenge to non-browser traffic. Open Input → Proxy configuration and make sure Apify Proxy is enabled with the RESIDENTIAL group selected — datacenter IPs are reliably blocked on this site.

This is not a soft recommendation: a production run on Apify's datacenter IPs was blocked on every retry and returned a single BLOCKED row, while the same input on residential IPs returned 30 reviews across two pages. The Actor recovers automatically from transient 403s by rotating sessions, but it cannot work around datacenter IPs entirely.

"PAGE_STRUCTURE_CHANGED" or "NO_REVIEWS_FOUND" error row instead of reviews. The Actor always checks the very first page of a product's reviews and refuses to end a run with a silent, unexplained empty result: if that first page loads successfully but none of the three extraction strategies (embedded React data, DOM, JSON-LD) find a single review, you get one of these two rows instead. NO_REVIEWS_FOUND means Capterra itself reports 0 total reviews for that product - nothing to scrape. PAGE_STRUCTURE_CHANGED means reviews were expected (the error message says how many, if known) but couldn't be extracted - Capterra likely changed its page layout; open an issue with the product URL.

A product's reviews stop after the first page, with no error row. This is expected once you're past the last page that has reviews - not a bug. (If pagination breaks entirely, e.g. Capterra changes its page-number query parameter, the Actor detects that a "next" page returned the exact same reviews as the one before it and stops rather than looping forever.)

A run was interrupted (timeout, manual abort, restart). Just run it again with the same input. Progress per product is saved to the key-value store, so already-finished products are skipped and in-progress ones resume from where they left off instead of re-scraping (and re-billing) from scratch.

Legal and ethical use

This Actor only requests pages Capterra's robots.txt allows (product and review pages) and does not crawl /search, login, or preview endpoints. You are responsible for how you use the scraped data — Capterra reviews are user-generated content; check Capterra's Terms of Use and applicable data-protection law (e.g. GDPR/CCPA) before storing or publishing reviewer-identifying fields, and avoid using this data to contact reviewers without consent.

FAQ

Can I scrape multiple products in one run? Yes — add multiple entries to startUrls and/or productSlugs; they're processed with bounded concurrency so one run can safely cover a whole competitor set.

Does sortBy change which reviews I get? No — it only reorders the reviews already collected. Capterra's robots.txt disallows crawling sorted result URLs (*?sort_options=), so this Actor never requests a specific sort order from the site itself; use minRating/maxRating if you need to narrow down which reviews are included.

Why do some fields come back null? Either Capterra didn't show that field for that particular review (e.g. not every reviewer gives a company size), or includeReviewerDetails is false. Fields are never undefined or omitted, so downstream schemas stay stable.

Can I get reviews for a product that isn't in my input list? Not directly — provide its URL or "{productId}/{productName}" slug. See Input above for why (robots.txt disallows site search).