A restaurant with 400 reviews and a 4.2-star average on Yelp can still have a real problem a single number won't show you: forty of those reviews mentioning slow service in the last two months, or a popular dish that started getting complaints right after a menu change. The average smooths all of that away. The review text is where it actually shows up, and for restaurants and hotels specifically, that text is split across two platforms with different reviewer bases — Yelp skews local and service-heavy, TripAdvisor skews travelers and trip-context.
If you're an agent (or building one) reading this rather than a human, the full machine-readable schema for every endpoint below lives at mindcase.co/skills.md.
What do star ratings hide that review text catches?
Timing and specificity. A 4.2-star average built from reviews over three years looks identical whether the business has been steady the whole time or started sliding six months ago — the number doesn't carry a date. It also collapses a hundred different complaints into one figure, so a business with consistently good food but recently bad service reads the same as one with the opposite problem. Reading the actual review text, filtered to recent dates, is the only way to catch a trend before it's fully priced into the average.
What does Yelp Business give you before you read a single review?
Structured signals that don't require reading anything. Yelp Business
API returns 25 fields per business, including
healthInspections, popularDishes, priceLevel, and reviewHighlights
— Yelp's own extracted summary of common review themes, sitting on top of
the raw reviews.
curl -X POST "https://api.mindcase.co/v1/data/yelp/business/run?wait=true" \
-H "Authorization: Bearer $MINDCASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"params": {
"urls": ["https://www.yelp.com/biz/katzs-delicatessen-new-york"]
}
}'
# $0.0005 per businesshealthInspections matters more here than it does for most business
categories — a restaurant's last inspection result is a hospitality-specific
risk signal with no equivalent in most other verticals, and it's a field
Google Maps doesn't return at all. popularDishes gives you a named
starting point for reading review text: instead of skimming every review
for whatever comes up, you know in advance which dishes get mentioned
often enough to search for.
How do you pull a restaurant's full review history from both platforms?
Say you're evaluating an independent restaurant before a franchise or acquisition conversation — not a chain with a known track record, a single location where the only real signal is what its customers have actually said. The business profile above tells you what Yelp thinks people talk about. Pulling the review text from both platforms tells you whether that holds up, and whether TripAdvisor's more travel-heavy audience saw something different.
How do you get the Yelp reviews?
curl -X POST "https://api.mindcase.co/v1/data/yelp/reviews/run?wait=true" \
-H "Authorization: Bearer $MINDCASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"params": {
"urls": "https://www.yelp.com/biz/katzs-delicatessen-new-york",
"maxReviews": 100
}
}'
# $0.00025 per reviewYelp Reviews returns 16 fields per review, including
publicReply — whether the owner responded, and what they said. A
business that replies to complaints reads differently than one that lets
them sit unanswered.
How do you get the TripAdvisor reviews?
curl -X POST "https://api.mindcase.co/v1/data/tripadvisor/reviews/run?wait=true" \
-H "Authorization: Bearer $MINDCASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"params": {
"urls": "https://www.tripadvisor.com/Restaurant_Review-g60763-d423981-Reviews-Katz_s_Delicatessen-New_York_City_New_York.html",
"maxResults": 100
}
}'
# $0.002 per reviewTripAdvisor Reviews returns 20 fields, and two
of them have no Yelp equivalent: tripType (business, couples, family,
solo, friends) and travelDate, separate from postedDate. A restaurant
that reads well in Yelp's mostly-local reviews but draws complaints in its
tripType: family TripAdvisor reviews specifically is telling you
something about which crowd it actually serves well — information a
combined star average would never surface.
Why not just use TripAdvisor's own review API?
Because it's built for something other than research at this scale.
TripAdvisor's official Content API caps review and photo endpoints at a
handful of items per location — around 5 reviews and 5 photos — regardless
of how many exist, and new developer accounts get roughly 5,000 free calls
a month with a daily search cap near 10,000 requests. That's workable for
displaying a small review widget on a business's own site. It isn't
workable for reading a full review history across multiple candidate
locations, which is the actual job here. TripAdvisor
Reviews has no per-location review cap —
maxResults set to 0 returns everything available for that place.
What can't you get from Mindcase?
Two cases.
A business with few reviews on one platform has little to read there. An independent restaurant that's well-established on Yelp but new to TripAdvisor will return a thin TripAdvisor result — true of any source, since neither Mindcase nor TripAdvisor itself can return reviews that haven't been written yet.
Sentiment scoring isn't built into either endpoint. Both return raw
review text, ratings, and Yelp's own reviewHighlights summary — not a
computed sentiment score or theme cluster across the pull. Turning a
hundred reviews into "service complaints trended up in the last two
months" is a step you run afterward, on the returned text.
Which endpoint should you use for which job?
| Endpoint | Input | Price | Best for |
|---|---|---|---|
| Yelp Business | Business URL or location + keyword | $0.0005 / business | Health inspections, popular dishes, price level, structured profile |
| Yelp Reviews | One business URL | $0.00025 / review | Local, service-heavy reviewer base; owner reply visibility |
| TripAdvisor Reviews | One place URL | $0.002 / review | Travel-context reviews; trip type and travel date fields |
FAQ
No. All three endpoints take a public business or place URL and run against public review data — no login or platform account required.
One Yelp Business call at $0.0005, 100 Yelp reviews at $0.00025 each ($0.025), and 100 TripAdvisor reviews at $0.002 each ($0.20) — about $0.23 total for a business profile plus 200 reviews across both platforms.
The per-row price reflects TripAdvisor's own access cost for this data, not a Mindcase markup — it's eight times Yelp's per-review price on this endpoint.
The field is returned per review (tripType), but the endpoint doesn't take it as a request filter — pull the reviews and filter the returned field yourself.
The healthInspections field returns what Yelp itself has published for that business. Coverage depends on Yelp's own data for the region and category, not on Mindcase.
