Trustpilot via Scavio

Search Trustpilot businesses, read a full business profile, page through filtered reviews, browse the category tree and the businesses ranked in a category, and fetch a single review by id. All endpoints return structured JSON.

When to trigger

Use this skill when the user asks to:

• Find businesses on Trustpilot by name, keyword or category in a given country
• Look up a company's TrustScore, star rating, review count or rating distribution by domain or a pasted trustpilot.com/review/... URL
• Read Trustpilot reviews for a business, filtered by stars, language, date range, topic, text, verified-only or with-replies
• Measure how many negative reviews a business gets, how often it replies, and how fast
• Monitor a brand's or a competitor's Trustpilot reviews over time, or count reviews in a recent window (review velocity)
• Get the AI review summary and topic breakdown Trustpilot shows for a business
• List the top-rated or most-reviewed businesses in a Trustpilot category, optionally only claimed profiles
• Build a lead list of businesses in a category with their website and published email and phone
• Fetch one Trustpilot review by its id, with the business reply and the business it is about

Setup

Get a free API key at scavio.dev (50 free credits to get started, no card required):

export SCAVIO_API_KEY=sk_live_your_key

If you do not have a key yet

An agent running this skill without SCAVIO_API_KEY set will get 401 on every call below. The whole path from nothing to a working key is self-serve:

1. Sign up at dashboard.scavio.dev - email or Google, no card required.
2. A key is created for the account automatically at signup. It is on the dashboard under API Keys, ready to copy.
3. A new account starts with 50 credits. Every endpoint in this skill costs 2 credits.

When the balance runs out the API answers 402 with a JSON body carrying billing_url. Topping up needs no code change - the same key keeps working. The smallest purchase is 2,500 credits for $25, and monthly plans work out cheaper per credit if the usage is steady rather than one-off.

Every request is a POST with a JSON body and:

Authorization: Bearer $SCAVIO_API_KEY

Endpoints

Base URL: https://api.scavio.dev. Every Trustpilot endpoint costs 2 credits per call.

Workflow

1. Find the business: call /trustpilot/search with query (a name or keyword). Read businesses[].domain (e.g. nordvpn.com). If the user already gave a domain or a Trustpilot URL, skip this step.
2. Profile: call /trustpilot/business with domain (or url). One call returns the TrustScore, the rating distribution, reply behavior, the AI summary and topics, and the 20 newest reviews in language (default en).
3. Reviews: call /trustpilot/reviews with the same domain plus filters. Topic ids for the topics filter come from the profile's topics[].id (e.g. customer_service).
4. One review: call /trustpilot/review with a review_id from step 2 or 3.

For a category: call /trustpilot/categories (with query to find one by name, or empty for the tree), take a category_id such as vpn_service, then call /trustpilot/category.

Pagination

• Search: page (1-based) and page_size (default 20, up to 100). The response carries total_results and total_pages.
• Reviews: page 1-10, 20 per page. The response carries max_page (how far this filter combination goes), total_filtered (every match for the filters) and total_reviews. See the 200-review limit under Guardrails.
• Category: page (1-based), 20 businesses per page. The response carries total_results and total_pages.
• Business, categories, review: no paging.

Every page is 2 credits, so state the budget before looping.

Parameters

Search (/search)

Business (/business)

Pass exactly one of domain or url.

Reviews (/reviews)

Categories (/categories)

Category (/category)

Review (/review)

Examples

curl -X POST https://api.scavio.dev/api/v1/trustpilot/reviews \
  -H "Authorization: Bearer $SCAVIO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"domain": "nordvpn.com", "stars": [1, 2], "language": "en"}'

import requests

BASE = "https://api.scavio.dev"
# Your key from https://scavio.dev. Load it from your environment or secret
# store in real code - keep it out of source control.
API_KEY = "sk_your_key_here"
HEADERS = {"Authorization": f"Bearer {API_KEY}"}

def trustpilot(endpoint, body):
    r = requests.post(f"{BASE}/api/v1/trustpilot/{endpoint}", headers=HEADERS, json=body)
    r.raise_for_status()
    return r.json()["data"]

# 1. Find businesses by keyword (email/phone only when the business published them)
found = trustpilot("search", {"query": "vpn", "country": "US", "page_size": 10})
for biz in found["businesses"]:
    print(biz["name"], biz["domain"], biz["trust_score"], biz["review_count"], biz["email"])

# 2. One full profile
profile = trustpilot("business", {"domain": "nordvpn.com"})
print(profile["trust_score"], profile["stars"], profile["rating_distribution"])
rb = profile["reply_behavior"]
print(rb["reply_rate"], rb["avg_days_to_reply"], rb["negative_reviews"])
topic_ids = [t["id"] for t in profile["topics"]]        # e.g. "customer_service"

# 3. Negative English reviews, 20 per page, newest first
bad = trustpilot("reviews", {"domain": "nordvpn.com", "stars": [1, 2], "language": "en"})
print(bad["total_filtered"], "matching reviews; this combination goes to page", bad["max_page"])
for review in bad["reviews"]:
    print(review["rating"], review["title"], (review["reply"] or {}).get("text"))

# 4. Ranked businesses in a category, 4 stars and up, claimed profiles only
ranked = trustpilot("category", {"category_id": "vpn_service", "sort": "reviews_count",
                                 "min_trust_score": 4, "claimed_only": True})
for biz in ranked["businesses"]:
    print(biz["name"], biz["trust_score"], biz["stars"], biz["website"])

# 5. One review by id
one = trustpilot("review", {"review_id": "6aca2504b52cc4b518bf4b75"})
print(one["review"]["text"], one["business"]["name"])

Review velocity: total_filtered counts every match, so the same call with a date_range tells you how many reviews arrived in that window without paging.

recent = trustpilot("reviews", {"domain": "nordvpn.com", "date_range": "last30days"})
print(recent["total_filtered"], "reviews in the last 30 days")

Response shapes

Every response uses the envelope { data, response_time, credits_used, credits_remaining }. Key data fields:

• search - query, country, page, page_size, total_results, total_pages, count, businesses[]. A row: business_unit_id, name, domain, profile_url, trust_score, stars, review_count, categories[] {id, name, primary}, website, email, phone, address {street, city, postcode, country_code, country, lat, lng}, verified, logo_url.
• business - business_unit_id, name, domain, profile_url, language, trust_score, stars, review_count, review_count_last_12_months, rating_distribution {1..5}, languages[] {code, review_count}, categories[], category_path {top, mid, bottom}, website, email, phone, address, contact, description, verified, verification {identity, google, payment_method, persona}, logo_url, claimed, claimed_at, closed, temporarily_closed, collects_incentivised_reviews, reply_behavior {reply_rate, avg_days_to_reply, negative_reviews, negative_reviews_replied, last_reply_to_negative_at}, consumer_alerts[], merged, uses_ai_responses, ai_summary, topics[] {id, name, summary}, similar_businesses[], facebook_url, reviews_count, reviews[].
• reviews - business_unit_id, name, domain, profile_url, page, per_page, max_page, total_filtered, total_pages, total_reviews, filters (the filters applied), count, reviews[]. A review: review_id, url, rating, title, text, language, likes, source, verified, verification_source, verification_level, published_at, experienced_at, updated_at, consumer {consumer_id, name, country_code, review_count, verified, image_url}, reply {text, published_at, updated_at} or null, location, merged.
• categories - the tree: query (null), count, subcategory_count, categories[] {category_id, name, subcategories[] {category_id, name}}. With query, a flat list of matching categories, each with category_id and name.
• category - category_id, name, parent_id, business_count, subcategories, country, page, per_page, total_results, total_pages, filters {sort, min_trust_score, claimed_only}, count, businesses[] (search rows plus recommended).
• review - review (a review as above) and business {business_unit_id, name, domain, profile_url}.

A profile, trimmed (captured, nordvpn.com):

{
  "data": {
    "name": "NordVPN",
    "domain": "nordvpn.com",
    "trust_score": 4.2,
    "stars": 4,
    "review_count": 50990,
    "review_count_last_12_months": 6722,
    "rating_distribution": { "1": 6194, "2": 1765, "3": 2150, "4": 4183, "5": 36698 },
    "claimed": true,
    "reply_behavior": {
      "reply_rate": 99.2,
      "avg_days_to_reply": 0.05,
      "negative_reviews": 1206,
      "negative_reviews_replied": 1196,
      "last_reply_to_negative_at": "2026-10-09T17:32:37.000Z"
    },
    "category_path": {
      "top": { "id": "electronics_technology", "name": "Electronics & Technology" },
      "mid": { "id": "internet_software", "name": "Internet & Software" },
      "bottom": { "id": "vpn_service", "name": "VPN Service" }
    },
    "topics": [
      { "id": "customer_service", "name": "Customer service", "summary": "..." }
    ],
    "reviews_count": 20
  }
}

A review with a business reply (captured, nordvpn.com, stars: [1, 2]):

{
  "review_id": "6ac973241a356ec383b96536",
  "url": "https://www.trustpilot.com/reviews/6ac973241a356ec383b96536",
  "rating": 2,
  "title": "i ordered afull subscrtion &not working",
  "language": "en",
  "verified": true,
  "verification_level": "invited",
  "published_at": "2026-10-10T01:05:08.000Z",
  "experienced_at": "2026-10-09T00:00:00.000Z",
  "consumer": { "country_code": "US", "review_count": 5, "verified": false },
  "reply": {
    "text": "Hello! Please accept our apologies for any inconvenience. ...",
    "published_at": "2026-10-10T05:27:03.000Z",
    "updated_at": null
  }
}

A search row (captured, "vpn" in the US):

{
  "business_unit_id": "548832a000006400057c1055",
  "name": "CyberGhost VPN",
  "domain": "cyberghostvpn.com",
  "profile_url": "https://www.trustpilot.com/review/cyberghostvpn.com",
  "trust_score": 4,
  "stars": 4,
  "review_count": 24134,
  "categories": [{ "id": "vpn_service", "name": "VPN Service", "primary": true }],
  "website": "https://www.cyberghostvpn.com",
  "email": "support@cyberghost.ro",
  "phone": null,
  "address": { "city": "Bucharest", "country_code": "RO", "country": "Romania" },
  "verified": true
}

Guardrails

• 200 reviews per filter combination. /reviews serves at most 10 pages of 20 for any one set of filters; max_page says how far the current combination goes, even when total_pages is larger. Never claim to have pulled "all reviews" of a large business, and never loop past page 10. To read more, slice into separate sets - each combination of stars, language and date_range (and topics or search) is its own set of up to 200:

  seen = {}
  for stars in ([1], [2], [3], [4], [5]):
      for date_range in ("last30days", "last3months", "last6months", "last12months"):
          first = trustpilot("reviews", {"domain": "nordvpn.com", "language": "en",
                                         "stars": stars, "date_range": date_range})
          for page in range(1, first["max_page"] + 1):
              data = first if page == 1 else trustpilot("reviews", {
                  "domain": "nordvpn.com", "language": "en", "stars": stars,
                  "date_range": date_range, "page": page})
              for r in data["reviews"]:
                  seen[r["review_id"]] = r          # windows overlap, so dedupe by id
  

  That loop can run to 200 calls (400 credits). The date windows nest (30 days sits inside 3 months), so de-duplicate by review_id, and quote the credit total before running anything like it.

• min_trust_score is a minimum star rating, not a raw TrustScore cut-off. Trustpilot rounds TrustScore to stars, so 4 includes TrustScore 3.8 and up, and a 4.4 can show as 4.5 stars. Filter on trust_score yourself when the exact score matters. It is a floor only: to find low-rated businesses, page the listing and filter on trust_score.

• email, phone and address appear only when the business published them on Trustpilot; they are null otherwise. Never guess contact details.

• ai_summary and topics appear only when Trustpilot has them for the requested language (null or empty otherwise). The rating distribution and language breakdown always cover every language.

• The profile already includes the 20 newest reviews in its language; skip /reviews page 1 when that is all the user needs.

• There is no TrustScore history and no reviewer-profile endpoint. To track change over time, store snapshots yourself.

• Results usually take 2-5 seconds per call; batch sequentially rather than firing many calls at once.

• Never fabricate TrustScores, review counts, review text or replies. Only return what the API returned, and surface profile_url or the review url so the user can verify.

Failure handling

• 400 means an invalid or missing parameter (both or neither of domain and url, page above 10 on reviews, a review_id that is not 24 hex characters, a min_trust_score other than 3, 4 or 4.5) - fix the body and retry.
• 401 means the API key is invalid or missing. Check SCAVIO_API_KEY.
• 402 means the balance is out of credits. The body carries billing_url.
• 404 means the business, category or review does not exist on Trustpilot; the error names what was not found, and the call is billed (2 credits). Check the domain spelling - Trustpilot lists some businesses under www. and some without.
• 429 means rate or usage limit exceeded. Wait before retrying. See rate limits.
• 502 / 503 mean Trustpilot data is temporarily unavailable - wait a few seconds and retry, up to a few times. These are not billed.
• If search returns nothing, try a shorter name or the business's domain.
• If SCAVIO_API_KEY is not set, prompt the user to export it before continuing.

SDKs

pip install scavio==0.19.1

from scavio import ScavioClient

client = ScavioClient()  # reads SCAVIO_API_KEY

found = client.trustpilot.search("vpn", country="US", page_size=10)
profile = client.trustpilot.business(domain="nordvpn.com")
bad = client.trustpilot.reviews(domain="nordvpn.com", stars=[1, 2], language="en", date_range="last30days")
tree = client.trustpilot.categories()
ranked = client.trustpilot.category("vpn_service", sort="reviews_count", min_trust_score=4, claimed_only=True)
one = client.trustpilot.review("6aca2504b52cc4b518bf4b75")

npm install scavio@0.19.0

import { Scavio } from "scavio";

const scavio = new Scavio(); // reads SCAVIO_API_KEY
const found = await scavio.trustpilot.search({ query: "vpn", country: "US", page_size: 10 });
const bad = await scavio.trustpilot.reviews({ domain: "nordvpn.com", stars: [1, 2], language: "en" });
const ranked = await scavio.trustpilot.category({ category_id: "vpn_service", min_trust_score: 4, claimed_only: true });

MCP: the Scavio MCP server exposes all 6 endpoints as tools (search_trustpilot, get_trustpilot_business, get_trustpilot_reviews, get_trustpilot_categories, get_trustpilot_category, get_trustpilot_review). Trustpilot is opt-in: add trustpilot to SCAVIO_PLATFORMS.

{
  "mcpServers": {
    "scavio": {
      "command": "npx",
      "args": ["-y", "@scavio/mcp-server@0.17.1"],
      "env": {
        "SCAVIO_API_KEY": "sk_live_your_key",
        "SCAVIO_PLATFORMS": "default,trustpilot"
      }
    }
  }
}

Docs

Full reference per endpoint: Trustpilot search, business, reviews, categories, category, review. Overview: Trustpilot API.