G2 Scraper

Collect structured G2 product profiles, software reviews, analytics, facet values, and product discussions. Use product URLs or slugs, or find products by keyword, category, or vendor.

Review intelligence

Area	Available data
Review text	Title, structured pros, cons, recommendations, benefits, full review text, and matching search snippets
Dimension ratings	Overall star value plus ease of use, setup, administration, support, requirements, security, and business ratings when G2 supplies them
Reviewer context	Company segment, industry, role, region, review source, contract length, and discount fields
Themes and switching	Loved and disliked themes, products switched from, switching reasons, and nested switching themes
Product analysis	Rating distributions, percentiles, trends, demographics, top themes, and switching-reason counts

Choose a collection mode

Goal	Main input	Dataset output
Collect product reviews	products, search, category, or vendor	default and products
Search review text	reviewSearch, with an optional product selector	default
Find similar reviews	similarTo	default
Fetch known reviews	reviewIds	default
Collect product analytics	products with analytics: true	products and analytics
Compare products	compare	products and analytics
Add valid filter values to a run	enumFields	metadata, followed by the selected main output

enumFields is additive. It writes the requested filter values to metadata, then the selected review, product, analytics, or comparison job continues. If you provide only enumFields, the run finishes after it writes metadata.

Collect reviews for one product

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

Use maxReviewsPerProduct: 0 only when you want all available matching reviews.

Find products before you collect reviews

Find products by keyword:

{
  "search": "project management",
  "maxProducts": 5,
  "maxReviewsPerProduct": 20
}

Browse a category:

{
  "category": "CRM",
  "maxProducts": 10,
  "maxReviewsPerProduct": 20
}

Browse products from one vendor:

{
  "vendor": "Salesforce",
  "maxProducts": 5,
  "maxReviewsPerProduct": 10
}

Filter product reviews

This example keeps low-rating enterprise reviews from the selected date:

{
  "products": ["slack"],
  "dateFrom": "2025-01-01",
  "ratings": ["1", "2"],
  "companySegment": ["enterprise"],
  "maxReviewsPerProduct": 100
}

Facet filters accept G2's facet values. First, list the values in the metadata Dataset:

{
  "enumFields": ["source_type", "response_type"]
}

Then use a listed value in a product review run:

{
  "products": ["slack"],
  "primaryRegions": ["Americas"],
  "sourceTypes": ["organic"],
  "responseTypes": ["video"],
  "maxReviewsPerProduct": 50
}

Search review text

Use a product selector to search within selected products:

{
  "products": ["slack"],
  "reviewSearch": "customer support",
  "includeHighlights": true
}

Omit the product selector to search reviews across products:

{
  "reviewSearch": "customer support",
  "maxReviews": 50,
  "collapseByProduct": true
}

Analyse or compare products

{
  "products": ["slack"],
  "analytics": true
}

{
  "compare": ["slack", "clickup", "jira"]
}

Comparison mode writes one product record and one analytics record for each resolved product. It does not create a separate comparison record.

Find known or similar reviews

Find reviews that are similar to supplied text:

{
  "similarTo": "The API was difficult to integrate",
  "maxReviews": 20
}

Fetch known review IDs:

{
  "reviewIds": ["13038740"]
}

Add product discussions

Set includeDiscussions in product review mode. Discussion records go to the discussions Dataset.

{
  "products": ["slack"],
  "maxReviewsPerProduct": 10,
  "includeDiscussions": true
}

Input

Field	Type	Default	Description
products	string[]		Product URLs, product slugs, product names, or G2 category URLs
maxReviewsPerProduct	integer	20	Maximum reviews for each product; 0 removes this limit
sortBy	string	DATE_DESC	DATE_DESC, DATE_ASC, RATING_DESC, RATING_ASC, HELPFUL_DESC, G2_DEFAULT
ratings	string[]	["1","2","3","4","5"]	1, 2, 3, 4, 5
companySegment	string[]		small_business, mid_market, enterprise
dateFrom	string		Lower date bound; use YYYY-MM-DD
dateTo	string		Upper date bound; use YYYY-MM-DD
primaryRegions	string[]		Americas, Asia Pacific, EMEA
sourceTypes	string[]		Collection-source facets; use enumFields to list available values
responseTypes	string[]		Review-medium facets, such as text or video
loveThemes	string[]		Themes that reviewers liked
hateThemes	string[]		Themes that reviewers disliked
switchingThemes	string[]		Reasons that reviewers switched products
discounts	string[]		Reported discount facets
reviewSources	string[]		Review-source channel facets
reviewSearch	string		Search review text; without a product selector, the search is global
search	string		Find products by keyword
category	string		Find products in a G2 category
vendor	string		Find products by vendor name
maxProducts	integer	10	Maximum products in search, category, or vendor mode; 0 removes this limit
compare	string[]		Product URLs, slugs, or names for product and analytics output
analytics	boolean	false	Write analytics instead of individual reviews for resolved product selectors
enumFields	string[]		source_type, response_type, love_theme_humanized, hate_theme_humanized, switching_theme_humanized, discount, review_source
similarTo	string		Find reviews that are similar to this text
reviewIds	string[]		Fetch reviews by G2 review ID or review page URL
maxReviews	integer	100	Maximum global or similar-search reviews; 0 removes the global-search limit, while Similar search stays limited to 1,000
collapseByProduct	boolean	false	Keep one review per product in global or Similar search; global search then returns at most 10,000 products
includeDiscussions	boolean	false	Also collect discussions in product review mode; ignored in analytics mode
includeHighlights	boolean	true	Include matching snippets when reviewSearch is set
proxyConfiguration	object	{"useApifyProxy":true}	Connection settings passed to Apify

The Actor requires at least one selector. enumFields always runs first and does not replace the main job. Of the dedicated main selectors, the Actor uses reviewIds before similarTo, and similarTo before compare, when more than one is present. A reviewSearch without a product selector runs globally. With products, search, category, or vendor, it filters the selected products instead. Repeated product selectors are preserved and can produce repeated records.

ratings, companySegment, date, region, facet, and reviewSearch filters apply to product reviews, global search, Similar search, analytics, and comparison. They do not apply to review-ID or facet-discovery modes. sortBy controls product and global review searches.

The Actor uses dateFrom and dateTo as supplied. It does not reject unsupported date formats or compare their order. Use YYYY-MM-DD, and make sure dateFrom is not later than dateTo.

The next example shows valid syntax for every input field. It is a field reference, not a recommended combination of modes.

Output

The Actor can write to five Datasets. Open the Dataset named in the table for the selected mode.

Dataset	When it is used	Record type
default	Product review, global search, Similar search, or review-ID mode	Review
products	Product, analytics, or comparison mode	Product
analytics	Analytics or comparison mode	Analytics
discussions	Product review mode with includeDiscussions	Discussion
metadata	Facet discovery mode	Facet values

Unavailable optional review values are null. Missing text can be an empty string. Lists can be empty.

Review example

This selected-field projection contains point-in-time values dated 28 June 2026. Reviewer identity and other fields are omitted. Use the field reference for the complete record.

{
  "id": "13038740",
  "productId": 3437,
  "productName": "Slack",
  "reviewUrl": "https://www.g2.com/products/slack/reviews/slack-review-13038740",
  "title": "One-Click Integrations That Just Work for slack",
  "pros": "Integration with other apps has to be 1 click",
  "cons": "UI when you send the messages. The typing window should be adjustable",
  "recommendations": null,
  "benefits": "connect and chatting with the team",
  "responseType": "text",
  "nps": 8,
  "npsScore": 4,
  "submittedAt": "2026-06-26T17:10:48.781-05:00",
  "publishedAt": "2026-06-26T19:30:00.900-05:00",
  "status": "approved",
  "isPublic": true,
  "highlightedSnippets": null,
  "scrapedAt": "2026-06-28T06:00:34.982Z"
}

Review field reference

All fields in this table are present. A nullable field has the JSON value null when G2 has no value for it.

Group	JSON field contract
Identity	id and productName: string; reviewUrl: string or null; productId: number
Text	title, rawText, and responseType: string; pros, cons, recommendations, and benefits: string or null
Ratings	nps, npsScore, easeOfUse, easeOfSetup, easeOfAdmin, easeOfDoingBusinessWith, meetsRequirements, qualityOfSupport, securityConfidence, securityImportance, and direction: number or null
Reviewer	userName: string; country, region, primaryRegion, companySegmentLabel, industryLabel, and roleLabel: string or null; companySegment, industry, and role: number or null
Sentiment	loveTheme, loveThemeHumanized, loveThemeParentHumanized, hateTheme, hateThemeHumanized, and hateThemeParentHumanized: string or null
Switching	switchedFromOtherProduct and switchedReason: string or null; switchedFromProductIds: number array or null; switchingThemes: string array or null; switchingThemesNested: object array or null, with string name and humanizedName in each item
Publication	submittedAt and status: string; publishedAt, reviewSource, sourceType, contractLength, and discount: string or null; helpful: number; englishAvailable, isPublic, published, and productPaidProfile: boolean; price: number or null
Product context	productIds and categoryIds: number array or null
Additional values	defaultSort, serviceLocation, contractLengthWeight, and discountWeight: number or null; templateKeyword, videoReviewSubmittedAt, and userUpdatedAt: string or null; editionIds and expertiseIds: string array or null; everApproved, isInLimbo, read, and reportable: boolean
Search and collection	highlightedSnippets: string array or null; scrapedAt: string

Reviewer job titles and avatar URLs are not available. industry and role are numeric G2 codes, and their label fields are present only for mapped values. They are output values, not filters. country and region are also output values; use primaryRegions for geography filtering.

Product field reference

All product fields are present.

Group	JSON field contract
Identity	name, slug, description, productUrl, and vendorName: string
Ratings	starRating, g2Rating, and mozPageAuthority: number or null; averageRating: string; reviewCount: number
Classification	categoryNames, featureKeywords, and tagNames: string array; categoryIds: number array
Images	thumbImageUrl and faviconUrl: string or null
Other values	type, uuid, and scrapedAt: string; isConsumer and isFeatured: boolean

Analytics field reference

All analytics fields are present. Arrays can be empty.

Group	JSON field contract
Identity and totals	productId, totalReviews, and totalHelpfulVotes: number; productName, productSlug, and scrapedAt: string
Rating statistics	nps, easeOfUse, easeOfSetup, easeOfAdmin, qualityOfSupport, meetsRequirements, easeOfDoingBusinessWith, securityConfidence, securityImportance, and direction: objects; each has number-or-null avg, stdDev, min, and max, plus number count
Distribution	ratingDistribution: object whose rating keys map to number counts; npsPercentiles: object with number p10, p25, p50, p75, and p90
Counts	uniqueCountries and uniqueIndustries: number
Top themes	topLoveThemes and topHateThemes: object arrays; each item has string theme, number count, and object-array sampleReviews; each sample has string title and userName, and number npsScore
Unique themes	uniquelyLoved and uniquelyHated: object arrays; each item has string theme and number score and count
Switching	switchingReasons: object array; each item has string reason and number count
Demographics	bySegment and byIndustry: object arrays with number code, string-or-null label, and number count; byRegion: object array with string region and number count; bySourceType: object array with string type and number count
Trends	reviewsOverTime: object array; each item has string month and number count

This analytics projection contains point-in-time values. It omits the other fields listed above.

{
  "productId": 3437,
  "productName": "Slack",
  "productSlug": "slack",
  "totalReviews": 20201,
  "totalHelpfulVotes": 69,
  "nps": {
    "avg": 9.211870699470323,
    "stdDev": 1.1668658162744323,
    "min": 0,
    "max": 10,
    "count": 20201
  },
  "ratingDistribution": {
    "1": 40,
    "2": 93,
    "3": 468,
    "4": 3646,
    "5": 15954
  },
  "npsPercentiles": {
    "p10": 8,
    "p25": 9,
    "p50": 10,
    "p75": 10,
    "p90": 10
  },
  "uniqueCountries": 140,
  "uniqueIndustries": 145,
  "reviewsOverTime": [
    { "month": "2014-02", "count": 1 },
    { "month": "2026-06", "count": 73 }
  ]
}

Discussion and facet-value fields

Discussion records have this contract:

Group	JSON field contract
Identity	id: string; productId and categoryId: number or null; productName and categoryName: string or null
Content	title and content: string; tagNames: string array
Counts	commentsCount, votesUp, votesTotal, and helpful: number
Flags	isPublic, unanswered, seeded, pinned, and vendorQuestion: boolean
Dates	createdAt, updatedAt, and scrapedAt: string

Some discussions contain a title without content. In a facet-value record, field is a string, type is the string enum_values, values is a string array, and count is a number equal to values.length.

This facet projection contains point-in-time values. The values array is shortened; count describes the full array.

{
  "field": "source_type",
  "type": "enum_values",
  "values": ["organic", "user_profile", "vendor"],
  "count": 21
}

Mode and result behaviour

• Product review and analytics modes return public, approved reviews. A requested review ID is omitted when it does not resolve to such a review, and the log names it. Review-ID records do not have a guaranteed order.
• A product name, such as Microsoft Teams, finds the G2 product with that exact name. The log shows the slug that it found.
• A run that reaches its maximum cost per run stops early and keeps the results that it charged for. The run status message tells you so.
• Similar search returns at most 1,000 reviews and keeps relevance order.
• Global search with collapseByProduct returns at most 10,000 products. The run log tells you when it stops at this limit.
• G2 search can return reviews of products that G2 no longer lists. G2 has no page for these reviews, so their reviewUrl is null.
• maxReviewsPerProduct: 0, maxProducts: 0, or global maxReviews: 0 can produce a long run.
• Product reviewCount is G2's total. It can be higher than the number of reviews that the Actor can collect.
• Unknown products, empty searches, and modes with no matches can finish with empty Datasets.
• If one product fails, completed sibling products remain available. A product can also keep partial output if G2 stops returning reviews early. If all requested product targets fail, the run completes with empty output.
• Some reviews do not have distinct pros, cons, or benefits. The Actor does not create missing text.
• rawText holds all answers of a review as one text, with a line break between answers. The Actor splits the text only when the number of parts fits the review form: 3 parts give pros, cons, and benefits. Until July 2022 the form also asked for recommendations to others, so 4 parts in a review submitted before August 2022 give pros, cons, recommendations, and benefits. Any other count means that an answer has its own line breaks, for example a list, so all four sections are null. rawText always has the full text.