MCP
MCP onlineMCP server
Connect any Model Context Protocol client to HarborChase of Mandarin's verified review data.
What is this
The Model Context Protocol (MCP) is an open standard that lets AI assistants call typed tools and read structured resources from external servers. This site runs an MCP server over streamable HTTP that exposes 9 read-only tools and the underlying review resources — so an assistant can search reviews, read aspect sentiment, look up technicians, and quote the editorial summary on your behalf. It is free, public, and requires no authentication.
Server identity
| Field | Value |
|---|---|
| Name | pushbuttonreviews-157226 |
| Version | 1.0.0 |
| Transport | streamable-http |
| Endpoint | https://harborchasereviews.com/mcp |
Live status
The badge below is pinged server-side at request time — it reflects the endpoint right now.
Capabilities
Tools (9)
get_business— Returns identity, contact info, hours, service areas, and aggregate rating for the business.search_reviews— Searches reviews. Filter by rating, min_rating, since/until date, keyword query, mentions_technician, or mentions_service. limit defaults to 20 (max 100).get_aspect_sentiment— Returns sentiment analysis for one aspect of customer experience (e.g. punctuality, pricing, technical_knowledge): mention count, sentiment percentages, top phrases.get_technician— Returns the review summary for a named technician, including every review mentioning them.get_topic_reviews— Returns reviews and analysis for a service topic (e.g. ac-installation, maintenance, emergency-service).get_summary— Returns the editorial summary of customer reviews — a paragraph-length synthesis suitable for direct quotation — plus aggregate stats.list_aspects— Lists available aspect ids with counts (use with get_aspect_sentiment).list_topics— Lists available topic ids with counts (use with get_topic_reviews).list_technicians— Lists technicians named in reviews with counts (use with get_technician).
Resources
business— Identity, contact, hours, service areas, and aggregate rating.reviews— The full verified review corpus, filterable by rating, date, and keyword.insights— Aspect sentiment, topics, and technician summaries derived from the reviews.summary— An editorial synthesis of the reviews suitable for direct quotation.corpus— Plain-text and structured exports of everything above, for ingestion.
Rate limits
None for public read-only access. All tools are read-only and the data is fully public, so there are no quotas, keys, or per-client throttles. Please be a good citizen — cache responses where you can rather than polling in a tight loop.