G2 Latest Reviews Monitor

Collects the latest reviews G2 publishes on each product's public RSS feed and returns them as flat rows, ready for JSON, CSV or Excel.

It is built for monitoring: run it on a schedule, keep an eye on what people are saying about your product and your competitors', and pull only what is new since last time.

How it works

The actor reads the public RSS feed G2 serves for a product:

https://www.g2.com/products/{product}/reviews.rss

That feed is open. In the checks run for this actor it answered without a login, without cookies, without a browser and without a proxy — including from a datacenter IP. Nothing is bypassed and no private page is touched: the actor asks for the same feed a feed reader would.

What it is not

This is a monitor, not an archive. G2's public feed is a snapshot of recent reviews with no way to page back through older ones, so:

• the default feed carries roughly the 25 most recent reviews of a product
• asking for more also reads the per-rating feeds and merges them, which brings in more recent reviews — but they overlap, so the number of distinct reviews you actually get depends on the product
• maxReviewsPerProduct accepts up to 125. That is a ceiling, not a promise: a quiet product may return a handful of reviews, and a busy one will not return its older history either
• if you need every review a product has ever received, this actor is the wrong tool

Input

One product

{ "products": ["slack"] }

A product URL works just as well; only the slug is kept.

{ "products": ["https://www.g2.com/products/slack/reviews"] }

Several products, filtered and incremental

{
  "products": ["slack", "asana", "https://www.g2.com/products/notion/reviews"],
  "maxReviewsPerProduct": 50,
  "ratings": [1, 1.5, 2],
  "publishedAfter": "2026-08-01T00:00:00Z",
  "knownReviewIds": ["13357798", "13357612"],
  "includeReviewerName": false,
  "maxTotalReviews": 500,
  "requestConcurrency": 3
}

Input fields

Field	Type	Default	Description
products	array of strings	— (required)	Product slugs or G2 product URLs, 1–100 entries. A product listed twice is collected once. The URL you type is never stored or logged; only the slug is kept.
maxReviewsPerProduct	integer	25	Reviews to keep per product, newest first. 1–125. Up to 25 reads the default feed only; above 25 also reads the per-rating feeds.
ratings	array of numbers	not set	Keep only these displayed star ratings: 0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5. Omit to keep every rating. An empty array is rejected.
publishedAfter	string	not set	ISO 8601 with a time zone, e.g. 2026-08-01T00:00:00Z. Keeps reviews published strictly after this moment.
knownReviewIds	array of strings	[]	Review IDs you already have; matching reviews are skipped and never charged. Up to 10,000 entries.
includeReviewerName	boolean	false	Off by default. Turn on only if you need the public display name and may store it.
maxTotalReviews	integer	1000	Ceiling on stored reviews for the whole run, so a run's price stays predictable. 1–10,000.
requestConcurrency	integer	3	Feed requests in flight at once across the whole run. 1–5.

Bad input is rejected before the run charges anything or fetches anything.

Star ratings are the stars on the review

You pick real star values. G2's own feed filter works in coarser bands, so the actor uses those bands only to fetch candidates and then re-checks every review against the rating it actually displays. Asking for 5 gives you 5-star reviews, not 4.5-star ones. A review whose rating cannot be read is left out when a rating filter is set — it is never rounded or guessed.

Incremental runs

Two fields let a scheduled run pick up where the last one stopped:

• publishedAfter — strictly greater than. Pass the newest publishedAt your last run returned and you continue from exactly there, with no repeat of the boundary review. A review with no publication date is skipped while this filter is set.
• knownReviewIds — pass the reviewId values you already stored. Send the most recent ones; the list is capped at 10,000.

Either one on its own works. Using both is fine.

Output

One flat row per review in the default dataset. The run summary — per-product status and the counts from every filtering stage — is stored separately under the OUTPUT key of the key-value store, so the dataset stays pure review rows.

Field	Type	Description
productSlug	string	Product the review belongs to.
productName	string | null	Product name from the feed.
productUrl	string	Canonical G2 reviews page.
reviewId	string | null	Numeric ID taken from the review URL. Pass these back as knownReviewIds.
reviewUrl	string	Canonical URL of the review.
title	string | null	Review headline.
rating	number | null	Stars shown on the review: 0–5 in steps of 0.5.
pros	string | null	What the reviewer likes best.
cons	string | null	What the reviewer dislikes.
problemsSolved	string | null	What problems the product solves for them.
publishedAt	string | null	ISO 8601 timestamp in UTC.
reviewerRole	string | null	Public role label, e.g. G2 User.
companySize	string | null	Public band, e.g. 2-10 employees.
reviewerName	string | null	Public display name. null unless includeReviewerName is on.
source	string	Always g2-rss.

About the nulls

A field is null when the feed did not publish it. The actor never fills a gap with a guess: an unreadable date stays null rather than becoming the collection time, a missing rating stays null rather than being inferred, and a review ID appears only when it can be read from the review URL. reviewerRole is usually present; companySize is published for most reviews but not all.

Every row is flat, so CSV and Excel exports are plain columns with no nested objects to unpack.

Pricing

Pay per event:

Event	Price
actor-start	$0.01 per run
review-saved	$0.002 per successfully stored review

The most a run can cost:

$0.01 + (reviews stored x $0.002)

Reviews stored	Maximum cost	Shown as
1	$0.012	$0.02
25	$0.06	$0.06
1,000	$2.01	$2.01

The actor rounds the ceiling it prints up to the whole cent, so the figure you see is never lower than what you pay. You are billed on the exact arithmetic, not on the rounded figure. The actor prints its own ceiling at the start of every run, worked out from maxReviewsPerProduct, the number of products and maxTotalReviews, before it fetches anything.

What you are not charged for

A review is charged only after the dataset has accepted it. You pay nothing for:

• reviews that arrived from more than one feed (duplicates)
• reviews excluded by your rating filter
• reviews excluded by publishedAfter
• reviews excluded by knownReviewIds
• products that returned no reviews
• products that could not be found
• reviews that failed to store
• any failed request or failed product

The start event is the exception: it is charged once per run, including a run that ends up storing no reviews at all. Keep maxTotalReviews in line with what you actually want so the ceiling stays where you expect it.

If storing succeeds but the charge does not go through, the reviews stay in your dataset and the run reports them as uncharged. The reverse — charging for something that was not stored — cannot happen.

Scheduling

A monitoring setup usually looks like this:

1. Run once with publishedAfter set to where you want to start.
2. Store the newest publishedAt you received.
3. Schedule the actor (Apify Console → Schedules, e.g. daily at 07:00) with that value in publishedAfter.
4. After each run, move publishedAfter forward to the newest publishedAt in the results.

If you prefer to track IDs, keep the reviewId values instead and pass the most recent of them as knownReviewIds. Either way, reviews you already have are skipped and not charged.

Run status

Each product ends in one of these states, reported in the run summary:

Status	Meaning
succeeded	Every planned feed answered and reviews were kept.
partial	Some feeds answered, others failed. What was collected is kept.
no_reviews	Feeds answered fine, but nothing survived your filters — or the product has no reviews. Not an error.
invalid_product	G2 has no such product.
blocked	The request was refused.
rate_limited	Too many requests; the actor waited and retried.
temporary_error	A server or network problem, or the product ran out of time.
parse_error	The feed came back but could not be read as RSS.

One failing product does not stop the others. The run as a whole succeeds if every product succeeded or had no reviews, is partial if some worked, and fails only if every product failed.

Exporting

Results export straight to JSON, CSV, Excel, XML or HTML table from the Apify Console, or through the API. Every field is a plain column.

Limitations

Worth knowing before you rely on it:

• The feed is a snapshot. There is no paging back through older reviews.
• The default feed carries roughly the 25 most recent reviews. Merging the per-rating feeds adds more recent candidates, but they overlap heavily, so the number of distinct reviews varies by product.
• 125 per product is a ceiling, not an amount you should expect.
• Products are processed one at a time, in the order you list them, so run time grows with the number of products.
• A restarted run may repeat work. Within a single process the actor never stores or charges for the same review twice, but if the platform restarts the run, or a request's outcome is unknown, duplicates are possible. Use knownReviewIds or publishedAfter to control this yourself. Automatic state between runs is a candidate for a later version.
• knownReviewIds holds at most 10,000 entries. Send the most recent ones.
• The actor depends on G2's public feed. If its structure changes, parsing may need an update; watch for parse_error in the run summary.
• includeReviewerName puts personal data in your dataset. Turning it on makes you responsible for handling that data lawfully.
• You are responsible for your own compliance with G2's terms and any laws that apply to you.

Privacy

Reviewer display names are left out unless you ask for them. The run summary and the logs hold counts, HTTP statuses and response hashes only — never review text, review URLs, review IDs, reviewer names or the product URLs you typed.

Troubleshooting

A product returns invalid_product. Check the slug in the G2 URL: https://www.g2.com/products/<slug>/reviews. The actor takes only the slug, so a typo in it is the usual cause.

Fewer reviews than maxReviewsPerProduct. Expected. The feed only carries recent reviews, and the per-rating feeds overlap. Check collected and the per-stage counts in the run summary to see where reviews went.

Everything filtered out. Look at filteredOutByRating, filteredOutByDate and filteredOutByKnownId in the summary. A publishedAfter in the future, or a rating filter no recent review matches, empties the result.

rating is null. The feed did not show a rating for that review. The actor leaves it empty rather than guessing, and a rating filter excludes such reviews.

blocked or rate_limited. Lower requestConcurrency and try again later.

Duplicates across runs. Pass the reviewId values you already have as knownReviewIds, or move publishedAfter forward.

Support

Questions and issues: open an issue on the actor's page in Apify Console.

---

This actor is an independent tool. It is not affiliated with, endorsed by or connected to G2, and it does not use any G2 API.