Review Radar — Google Maps review alerts to Slack via MCP connector

Monitor Google Maps reviews for a portfolio of businesses, detect only what's new across runs (persistent state), and post Slack alerts for negative reviews through an Apify MCP connector — the Actor never sees a credential, only a connector ID.

Why this Actor is different

Most review scrapers are stateless: every run returns everything, and you diff by hand. Review Radar keeps a named key-value store of seen review IDs per place, so a scheduled run surfaces exactly the reviews published since the last run — the difference between a scraper and a monitor. The Slack side uses MCP connectors: you authorize Slack once in your Apify account, the Actor receives a connector ID at runtime, and the proxy enforces which tools it may call.

Use this tool when

• You run an agency and monitor reviews across a client portfolio (hotels, restaurants, local businesses)
• You want a Slack ping the moment a damaging review lands — not a weekly report
• You want a working example of persistent state (named KV store) + optional MCP connector in one Actor

Input

Field	Description
placeUrls	Google Maps /maps/place/... URLs to monitor
query	Alternative: a Maps search (first maxPlaces results)
maxPlaces	How many businesses (default 3, max 20)
maxReviewsPerPlace	Recent reviews scraped per business (default 10, max 50)
alertThreshold	New reviews at or below this rating trigger an alert (default 3)
slackConnector	Optional Slack MCP connector. Empty = dataset only
proxyConfiguration	Residential proxy strongly recommended
dryRun	Skip Slack alerts and skip state writes (testing)

Output

One dataset item per review: place, author, rating, text, date, reviewId, isNew, alerted, mapsUrl, scrapedAt.

Prerequisites

• An Apify account (free plan works)
• Optional: a Slack workspace authorized as an MCP connector (Console → Settings → Integrations)
• A residential proxy for the Maps scraping

Run it

1. Open the Actor in Apify Console (or apify push from this folder).
2. Provide place URLs (or a search query), set the threshold, pick your Slack connector.
3. Run once to seed the state (everything reports as new), then attach a daily schedule — subsequent runs only report fresh reviews.

How it works

1. Playwright loads each place page (consent wall handled, media blocked, 240s handler timeout).
2. Reviews are scraped (author, rating from aria-label, text, relative date) and deduplicated by stable review ID.
3. Each ID is diffed against the named store review-radar-state; new IDs are flagged and persisted.
4. New reviews at or below the threshold are posted to Slack via the MCP proxy (ACTOR_MCP_CONNECTOR_BASE_URL + run APIFY_TOKEN).

Terms

Public review data only. Google Maps scraping sits uneasily with Google's ToS — keep volumes polite (daily cadence, few places), weigh the risk for production use, and prefer official APIs where they cover your need.

AI-agent ready, API + MCP ready.