Connect

n8n

n8n is a workflow tool. It has an MCP node built in, so a workflow can call Scrapewise directly: search Google and Amazon, find local businesses, read a page, read your saved rows. It can then send the result to Google Sheets, Slack, a database or anything else n8n connects to.

No AI model is needed. The node calls one Scrapewise tool you choose, with the arguments you give it. You need n8n 1.80 or newer (Cloud or self-hosted).

{
  "body": {
    "queries": ["lego duplo"],
    "marketplace": "de",
    "pages": 1,
    "quote_only": false
  }
}

Errors you may see

Questions

The six templates

1. Local business lead list

Download: 01-google-maps-leads-to-google-sheets.json

Searches Google Maps for a trade in a place, drops the businesses you cannot act on, appends the rest to a sheet, and posts one Slack digest saying how many leads came in.

Set query, location, country, language and max_results in Search settings. One call returns up to 20 places.

Two filters run before the sheet, both set in Search settings:

So the sheet holds leads you can contact rather than everything Google listed. Set min_rating to 0 and min_reviews to 0 to keep them all. Notify new lead then posts how many leads came in and the first ten by name, category, rating and contact, so you do not have to open the sheet to know whether the run was worth it.

Google Maps holds no e-mail addresses, so none are returned. The Website column is the starting point if you need e-mail: send those websites to the Website contacts endpoint as a second step. The template itself does not do that step.

To see the price before you spend anything, set quote_only to true in Search settings. The workflow then writes one row holding the quote and your wallet balance instead of leads.

2. Competitor price monitor

Download: 02-competitor-price-monitor-with-matching.json

Reads a price-monitor group you already have in Scrapewise, groups the rows by the column that matches your product to the competitor’s (usually EAN), and keeps only the products where a competitor is cheaper than you. Writes the gaps to a sheet and sends one Slack message with the ten biggest.

This template does not scrape. It reads what your scrapers already collected, so set the scrapers up first - the build-scrapers skill does that for you, or do it in the portal.

In Monitor settings, put in:

If a column name is wrong the workflow stops and the error message lists the column names the group actually has, so you can copy the right one straight out of it.

Prices written as 1.234,56 and 1,234.56 are both read correctly.

3. Amazon price tracker

Download: 03-amazon-price-tracker-to-google-sheets.json

Every morning, searches Amazon for each of your keywords and appends what it sells, with the date. Leave it running and the sheet becomes a price history you can chart.

Set queries, marketplace and pages in Keywords and marketplace.

Marketplaces: com, co.uk, de, fr, it, es, ca, com.au, co.jp, in, com.br, com.mx, nl, sg, ae, sa, pl, se, com.tr, eg, com.be.

One keyword and one page answers straight away. Two or more run as a background job, so the template has an If node and a Wait node that pick the rows up with get_results. If a long list is not finished in time, raise the wait.

One answer carries one page of rows. If Get job results comes back with truncated: true and a nextPage, there are more rows waiting. The template saves the first page, which is plenty for a handful of keywords. For a long list, send that nextPage value back into Get job results as page and keep going until truncated is false.

4. Google rank tracker

Download: 04-google-serp-rank-tracker.json

Every morning, searches Google for each keyword, looks for your domain in the results, writes the position you hold - or not in top results - and tells you in Slack when a keyword slips or falls out altogether.

Set domain, keywords, country and language in Keywords and domain.

google_search with type web takes one keyword per call, so the template loops over your list one keyword at a time with a Loop Over Items node. Ten keywords a day costs €0.015.

The domain match ignores www., so scrapewise.ai matches https://www.scrapewise.ai/pricing.

Three Slack messages come out of it, and they answer different questions:

The two per-keyword messages are exclusive, so a keyword produces one of them or neither - never both.

alert_below ships as 20, in Keywords and domain. It is the position you stop caring about: 20 means you are told when a keyword falls past 20. Lower it to 5 to hear about smaller slips, or raise it past the results you ever see to silence Ranking warning entirely. It does not affect Rank alert - a drop-out is always reported - and the digest arrives either way.

The digest reads the rows back out of the sheet rather than the search answers, so it reports what was actually saved. It only counts rows carrying today’s date, so a sheet holding months of history still produces a one-day summary.

5. Google review watcher

Download: 05-google-maps-reviews-monitor.json

Every morning, reads the reviews of the places you list, saves them all to a sheet, and sends a Slack message for every review at or below your star threshold.

Create an empty group of its own in the portal for this workflow and put its id in Places to watch. The read step returns everything in the group, so a shared group would mix other data in.

Place URLs need the place’s data_id, and the link you copy out of the Google Maps address bar does not carry one. Run template 1 (or the find_local_businesses tool) for the business - every result has a data_id that looks like 0x4692935f7b0b17f5:0xb136062b9dd26adb. Put it in a link of this shape, one per place:

https://www.google.com/maps/place/?data_id=0x4692935f7b0b17f5:0xb136062b9dd26adb

Each run writes today’s reviews again, so a review appears once per run with that day’s date. Sort or pivot the sheet on Review id to see each review once.

Set alertAtOrBelowStars to the threshold you want - 3 means three stars and below get a Slack message.

force is on, so you get fresh reviews every day and every place is read and charged each run.

If the sheet comes out empty, raise the wait in Give it two minutes - a long list takes longer.

6. YouTube keyword watch

Download: 06-youtube-keyword-watch.json

Every Monday, searches YouTube for each of your keywords, saves what comes up, and posts a Slack message for each video published in the last week. Run it weekly and the sheet shows which videos are gaining on your topics, and which of them are yours.

Create an empty group of its own in the portal and put its id in Keywords to watch, same as template 5.

Put your keywords in plain words - the workflow turns each one into a YouTube search link itself. Add your own channel name as myChannel and the Mine column flags your videos.

The whole result set goes to the sheet; only the fresh part is worth interrupting you for, so Published this week? keeps the videos published less than eight days ago and New video alert posts each one with its channel, views, length and link. Everything older is already in the sheet and no message is sent.

Two things to know about that filter. It reads YouTube’s own published date, which arrives as wording like 3 days ago rather than a date, so the age is derived from that wording and a phrasing it cannot parse is treated as old rather than new. And because the schedule and the window are both about a week, a video posted a few hours before one Monday can still be inside the window on the next - expect the occasional repeat rather than a missed video.

The alert covers every channel, including yours. If you only want other people’s, add a condition on isMine being false alongside the one already in Published this week?.

Templates 5 and 6 both send their list under a name. If you add or remove a place or a keyword later, change that name too - Scrapewise refuses a different list sent under a name it already knows.

How the templates are put together

Useful if you want to change one, or copy the pattern into a workflow of your own.

Three details the Code nodes handle for you:

• The answer arrives in one of two shapes. Already parsed in structuredContent, or as a string in content[0].text, depending on your n8n version. Every Code node reads both.
• Scraped strings are wrapped. { "type": "scraped", "content": "…", "origin": "…" } marks text a stranger wrote on a web page. The Code nodes unwrap it to a plain string. Never let a later AI step treat it as an instruction - see MCP safety.
• Reading a group’s rows returns a page. The rows are under content, not rows.

All six templates point at https://mcp.scrapewise.ai/mcp. Template 2 only reads rows you already collected, so a Read key is enough for it; the other five scrape and need a Full key.

Five of the six also post to Slack, so they need a Slack OAuth credential on top of the Scrapewise one. Only template 3 runs without it.

Good to know

• Check the price first. Templates 1 and 3 have a quote_only setting: turn it on and the call returns the price and your wallet balance without scraping. Templates 5 and 6 answer with estimatedMicroCents before the run starts.
• Start with one. One keyword, one place, one product. The cost per run is then obvious and the schedule can be turned on afterwards.
• A retry does not charge twice. Add the same idempotency_key inside body and identical requests within 24 hours are charged once.
• Three jobs at a time. An account can have three background jobs running. A fourth is refused until one finishes.
• Change the schedule freely. Every schedule in these files is a plain n8n Schedule Trigger - open it and set whatever you want, or swap it for a webhook.

Last updated 4 November 2026