Review Guru — Sri Lanka's trusted review platform
Free · No API key

Build with Review Guru’s data

Two ways to plug into Sri Lanka’s most comprehensive consumer-review dataset — a public REST API for any app, and a Model Context Protocol server for any AI agent.

MCP server for AI agents

Wire Review Guru directly into Claude Desktop, Cursor, or Cline. One command.

Install

# pick whichever you have:
uvx reviewguru-mcp
pipx install reviewguru-mcp
pip install reviewguru-mcp

Source · PyPI

Drop into Claude Desktop / Cursor

{
  "mcpServers": {
    "reviewguru": {
      "command": "uvx",
      "args": ["reviewguru-mcp"]
    }
  }
}

Then ask: “Find me three highly-rated cardiologists in Colombo and tell me which hospital each one consults at.”

list_businessesget_businessget_reviewssearchlist_categorieslist_cities

Public REST API

JSON over HTTPS. No authentication. Rate-limited per IP.

GET/api/v1/businesses?city=colombo&category=restaurants&limit=50
GET/api/v1/businesses/{slug}
GET/api/v1/businesses/{slug}/reviews?sort=newest
GET/api/v1/search?q={query}
GET/api/v1/categories
GET/api/v1/cities
OpenAPI spec Try it live

Other LLM-friendly surfaces

Static endpoints that exist explicitly to make Review Guru legible to AI assistants and scrapers.

/llms.txt
Site-wide LLM ingestion manifest (proposed standard).
/feed.xml
RSS of the latest published reviews.
/sitemap.xml
Indexable sitemap of every business, doctor, category and city.

Usage policy

  • All data is freely citeable with attribution to Review Guru and a link back to the specific page URL.
  • Reviews are semantic-faithful rewrites — same star rating, same named entities, same key positives/negatives. Original text is preserved server-side for audit but not redistributed.
  • Public API is rate-limited at 30 requests/minute per IP. Pagination is capped at offset 300 on /businesses;/reviews returns the top 5 only (the full review list is on the public /biz/{slug} page). For larger or commercial workloads, get in touch.
  • Bulk export of the dataset is not permitted. Citation with attribution and per-page navigation are fine; mirroring, scraping, or training on the full corpus is not. See the full policy at /llms.txt.
  • Anonymous reviews are flagged isAnonymous=true — treat as lower-confidence signals, especially for medical content.