Google Maps Scraper Alternative - Local Business Leads by City

What does Places Database do? Search businesses → get email, phone and social links in one run

• You give: a city and a business type, e.g. Austin, US + dentist (or München, DE + Zahnarzt).
• You get: one row per business with name, category, address, coordinates, website, phone (in +E.164 form), role e-mail addresses (info@, contact@) and social links, from the open Overture Maps places dataset. Select + Contact enrichment to read the business website for more emails, phones and social profiles in the same run.
• You pay: $1 per 1,000 places returned, no start fee. Locations that are not found, runs that match nothing, filtered-out places and duplicates are free. 20 places cost $0.02.
• Good to know: it is a snapshot of an open dataset, not a live map. In big cities roughly half of the records were last updated more than a year ago; the Max source age option keeps only recent ones (see What to expect).

No map website is scraped, so there are no proxies, no CAPTCHAs and no blocking; a city takes under a minute to a few minutes.

How to use it: get a lead list

1. In Cities or areas, enter one city per line with its country code: Lisbon, PT, Austin, US.
2. In Categories, enter the business type: dentist, restaurant, lawyer, hotel... (English, Overture names such as dental_clinic, or common words in German, French, Spanish, Portuguese, Italian, Dutch, Chinese and Japanese). Leave empty for every business.
3. Set Max places (the run never returns or charges more). For website contacts in this run, tick Only places with a website and choose + Contact enrichment under Add contacts or company details. Choose Email only if a phone or social link alone should stay free.
4. Click Start, then export the dataset as CSV, Excel or JSON. The default overview includes contactEmails, contactPhones, contactSocials and the three freshness fields. The directory view retains the original overview columns; full contains the source and enrichment details.

What else it does

• 🏙️ City, box or country: Austin, US, a bounding box 13.30,52.48,13.45,52.56, or a country code. About 290 major cities use their whole urban area (Austin includes Round Rock and Pflugerville, Berlin includes Potsdam); every other place is cut out along its real boundary (looked up with OpenStreetMap Nominatim). Each row has its own city field, so you can filter to the city proper.
• 🏷️ Category filter: plain words (dentist, lawyer, car repair) are mapped to Overture categories such as dental_clinic, attorney_or_law_firm, automotive_repair; Overture names work directly. Plurals and spaces are fine, and a parent category (food_and_drink, health_and_medical) includes everything under it. If nothing matches, the free no_results row and the log list the most common categories in your area. A category the Actor cannot read (for example a word in a language it has no translation for) is refused instead of being ignored, so you never get, and pay for, every business in the city by mistake. Overture's plumbing category also contains plumbing supply shops, and lawyer includes every kind of legal service.
• 🔗 Chains last: when one name has many branches (Starbucks, Droga Raia), the first branch is listed with the distinct businesses and the other branches follow, so a short list is not ten copies of one chain. All branches are still returned when Max places allows.
• ✅ Quality gate: confidence of at least 0.85 by default (Overture's score that the place exists), places the source marks as permanently closed removed.
• 🕒 Freshness filter: Max source age in months (for example 12) keeps only places whose source record was updated recently. Older records are likelier to be closed or moved. The run summary counts how many returned places are older than 12 months.
• 🕒 Freshness per row: sourceUpdated is the date in the Overture source, retrievedAt is when this Actor read the record, and lastVerifiedAt is the recorded time of a successful website read by this Actor's contact enrichment. Company classification runs through a separate Actor that does not return a website-check time, so its lastVerifiedAt stays null. lastVerifiedAt: null means no verifiable website-check time is available; none of these dates proves the business is still operating.
• 💬 Clear answers when nothing matches: a city, category or filter with no data ends with a free no_results row that says why (no data in the area, below your confidence, the filters, a category name that does not exist with the most common ones nearby), instead of an empty table.
• 🧹 Duplicates removed: same name and address, name and phone, name and spot (about 11 m), or same website, postcode and name start. Duplicates are never charged.
• 📇 Business fields only: no person names are extracted. E-mails are role addresses only (info@, contact@, sales@...) unless you switch on personal addresses yourself.
• 🔎 Filters: only places with a website, phone or e-mail; per-city caps; best-scored places first.
• ☎️ Clean contact fields: phone numbers are converted to international form (+493031800580) when they are valid for the place's country; numbers that do not parse are kept as the source wrote them. Click-tracking parameters (utm_*, gclid, fbclid) are removed from websites.
• 🧾 Honest run summary: release used, candidates, duplicates dropped, how many rows have a website, phone or role e-mail.
• 💵 $1 per 1,000 places, down to $0.80 on Apify's Business plan and higher. No start fee. Locations that are not found, filtered-out places and duplicates are free.

What data do you get?

Output example

One row per business (key fields; the full view has more):

{
    "id": "eb8be1be-fb13-4309-84af-f2578e753841",
    "name": "Gragger Brot",
    "category": "bakery",
    "categoryPath": ["food_and_drink", "casual_eatery", "bakery"],
    "website": "http://www.gragger.de/",
    "phone": "+493081866477",
    "emails": ["marketing@gragger.de"],
    "contactEmails": ["marketing@gragger.de"],
    "contactPhones": ["+493081866477"],
    "contactSocials": [],
    "emailType": "role-only",
    "address": "Potsdamer Str. 107",
    "city": "Berlin",
    "postcode": "10785",
    "country": "DE",
    "latitude": 52.5,
    "longitude": 13.36,
    "confidence": 0.996,
    "sourceUpdated": "2026-09-14",
    "retrievedAt": "2026-10-21T08:00:00.000Z",
    "lastVerifiedAt": null,
    "release": "2026-09-23.1",
    "license": "CDLA-Permissive-2.0",
    "attribution": "© Overture Maps Foundation (places)"
}

The coordinates and timestamps are illustrative; real rows carry full precision and the run's actual timestamps. With enrichment, the row also has enrichedEmails, enrichedPhones, enrichedSocials, a non-null lastVerifiedAt after this Actor's contact reader successfully reads the website, and flat linkedIns, twitters, instagrams, facebooks, youtubes, tiktoks URL arrays where present. Company classification has no verifiable check timestamp, so lastVerifiedAt stays null.

What the data is not: Overture has no ratings, reviews or opening hours, and only about 1 in 5 places has an e-mail address. It is a business directory, not a review database. For contact details beyond the source data, switch on the optional enrichment below, which reads each business's own website.

Input example

The form is prefilled with dental clinics in Lisbon, capped at 20 places (about $0.02, done in under a minute). A bigger input:

{
    "locations": ["Lisbon, PT", "Porto, PT"],
    "categories": ["restaurant", "hotel"],
    "requireWebsite": true,
    "maxResults": 2000,
    "enrich": "contacts",
    "enrichRequire": "email"
}

Next step: from places to verified email leads

Places Database gives you the businesses and their websites. Three TidyTools Actors turn that list into e-mail leads, each reading the previous run's dataset (copy the dataset ID from the run's Storage tab, or use {{resource.defaultDatasetId}} in an Apify integration):

Step	Actor	What it adds	Price
1	Places Database (this Actor)	businesses, websites, phones	$1 per 1,000 places
2	Website Email Extractor for Google Maps Leads & Domain Lists	e-mails, phones and social links read from each website	$2 per 1,000 websites with a contact found
3	Email List Cleaner - Bulk Email Validator	syntax, domain/MX and disposable checks on every address	$0.50 per 1,000 e-mails
4	Google Maps Lead Enrichment - Company Enrichment	industry, business model, description, tech stack	$6 per 1,000 companies classified

Copy-paste inputs (paste into each Actor's JSON input; replace the IDs with your own runs' dataset IDs):

1. Places Database:

   { "locations": ["Austin, US"], "categories": ["dentist"], "requireWebsite": true, "maxResults": 500 }
   

2. Website Email Extractor, reading the websites of step 1 and keeping each place's id, name, phone and city:

   { "datasetId": "<dataset ID of step 1>", "datasetField": "website", "datasetKeepFields": ["id", "name", "phone", "city"] }
   

3. Email List Cleaner, checking every address found in step 2:

   { "datasetId": "<dataset ID of step 2>", "datasetField": "emails" }
   

4. Company enrichment (optional), classifying the companies of step 1:

   { "datasetId": "<dataset ID of step 1>", "datasetField": "website", "datasetKeepFields": ["id", "name"] }
   

For 500 dentists with a website, steps 1 and 2 cost at most $0.50 + $1.00; step 3 adds $0.0005 per address checked; step 4 adds at most $3.00. Want contacts in one run instead? Use Add contacts or company details below ($3 per 1,000 websites with a contact found).

Three ways to buy the result

Pick what you need with Add contacts or company details. The log prints the cost of each option before the run starts.

Option	What you get	Price
Directory only (default)	Name, category, address, coordinates, website, phone, role emails and socials from the open data	$1 per 1,000 places
+ Contact enrichment	Directory + emails, phones and social links read from each business's own website	+ $3 per 1,000 websites where a contact was found (with Email only: where an email was found)
+ Company classification	Directory + industry, business model, description, contacts and tech stack, by Company Website Enrichment	+ $6 per 1,000 companies classified, billed by that Actor

Worked example, 100 dental clinics with a website in Berlin (measured 2 Oct 2026): directory $0.10; with contact enrichment $0.10 + 88 websites x $0.003 = $0.36; 69 of the 100 had an email address a buyer would see (role addresses), 51 of them new compared with the source data.

How enrichment behaves:

• Source values are never overwritten. What a website adds goes into enrichedEmails, enrichedPhones, enrichedSocials and an enrichment object (status, pages read, whether the page loaded, where each email was found).
• Export one contact list per business. contactEmails, contactPhones and contactSocials deduplicate source and website values. The social platform arrays use the same names seen in many Maps exports; they contain URLs, not enriched profile statistics.
• "Contact found" means the website showed one, even when it is the same address the source data already had; you pay for the website read that found it (enrichment.status: "found"), not only for new addresses.
• A website shared by several places (chains, branches) is read and charged once; every place keeps its own id, so branches are never merged.
• No website in the source means the website is unknown, not that the business has none. Such places are not charged for enrichment.
• A website that loads does not prove the business still operates. websiteReachable only says the page opened.
• Only role addresses (info@, contact@...) are returned unless you switch on personal addresses, the same rule as the directory data.
• The website reader identifies itself and respects robots.txt. Its user agent is TidyToolsContacts/1.0 (+https://apify.com/tidytools) (TidyToolsSmartFetch from our servers), never a browser's. Pages a website's robots.txt disallows are not read; a website whose home page is disallowed gets enrichment.status: "robots_disallowed" and is not charged.
• Contact-enrichment spending is capped by Max websites to read and by your run's maximum charge. Unreachable websites and websites with nothing found are free.
• Company classification runs as separate Company Website Enrichment runs, billed to you by that Actor. They are limited by Max websites to read (each sub-run gets its own maximum charge of $0.006 per website it is given), but they are not counted in this run's maximum charge.
• Cheaper for big lists: run the directory first, then give its dataset to Website Email Extractor ($2 per 1,000 websites with a contact), as in the next step section.

What to expect from the source data

Measured on 600 random places (Berlin, Austin, Lisbon; restaurants and dental clinics; 100 each; Overture 2026-09-23.1):

• Website field present: 87% of places (restaurants 69-88%, dental clinics 91-94%). The rest have an unknown website.

• Duplicates removed: 0-3.5% of candidates.

• Email in the source data: 17% of places (11-30% by city and category). Phone, website or social link: 97%.

• Home page opened for 80% of the listed websites.

• After reading the websites: the share of places with an email address rises from 17% to 53% when personal addresses are allowed, and to 33% with role addresses only.

• Source date: it depends on the city and the business type. In the sample above 87% of the records carry a sourceUpdated date in September 2026, but 500 lawyers and accountants in London (10 Oct 2026) were 45% updated within the last 12 months and 55% older, many from 2014-2017 (Microsoft source), while Nairobi, Taipei and Cairo samples were almost all from 2026. Restricting to the last 12 months in London kept 1,117 of 3,463 candidates and raised the share with an e-mail from 34% to 61% in the first 300 rows. It is the date of the source dataset's record, not a check that the business is open. About 1 in 9 listed websites did not resolve in DNS and about 1 in 10 returned an error or timed out; that points at stale entries but does not prove a closure.

These are samples, not guarantees: your city and category will differ.

Use with AI agents (MCP)

Connect Apify's MCP server (https://mcp.apify.com?tools=tidytools/places-database) to Claude, Cursor or any MCP client, then ask e.g. "list 200 dental clinics with a website in Lisbon". Minimal input:

{ "locations": ["Lisbon, PT"], "categories": ["dental_clinic"], "maxResults": 200 }

Problems come back as free rows with an errorType (invalid_input, not_found, no_results, failed) and a plain explanation; the run still ends normally. They are not charged.

Pricing

$1 per 1,000 places (place-returned event), $0.90 on Silver, $0.80 on Gold and above. No start fee. You pay only for places that are returned. Optional contact enrichment adds $3 per 1,000 websites with a contact found (place-enriched; $2.70 on Silver, $2.40 on Gold and above); company classification is billed by Company Website Enrichment. A run's worst case is shown in the log before it starts and is capped by your Max places and your run's maximum charge; company classification sub-runs are billed separately and are not counted in that maximum.

Limitations

• No ratings, reviews, opening hours or photos. Overture is a business directory; use a map scraper if you need those.
• E-mail coverage is low in the source (about 1 in 5 places). Contact enrichment raises it by reading each business's website, at an extra cost.
• A snapshot, not a live check. Some places have closed or moved since the source was updated; confidence and sourceUpdated help you judge.
• Category names are Overture's. Common words are translated (about 50 business types, in English and several other languages); anything else must be an Overture category name. A category that does not exist returns nothing (a free no_results row suggests the most common ones in your area), and a word the Actor cannot read at all ends the run with a free invalid_input row before anything is charged.
• Cities without a country code are guesses. Springfield becomes the best-known Springfield (Illinois); the run summary and log name the place used. Write Springfield, US or another country code to be sure. A US state after the comma (Springfield, Illinois) is read as the United States and is not used to pick the city: use a box to choose between cities of the same name.
• The source is older in some places. The share of records not updated for a year varies from almost none (Nairobi, Taipei) to more than half (London lawyers and accountants); use Max source age if freshness matters more than volume.
• Major cities mean the urban area. For about 290 big cities the search box covers the whole urban area, so nearby towns appear; use the city field or a smaller box to narrow it.
• Chains have many rows. Every branch is its own place with its own id; branches are listed after the distinct businesses but are not merged.
• Whole countries are slow. Add categories and a Max places limit; very large areas may need several runs.

Responsible use

This Actor returns business information. Please use it responsibly:

• Personal data. Some businesses are one-person practices, and their phone number or address is then personal data. This Actor does not extract person names and returns only role e-mail addresses by default. If you switch on Also return personal email addresses, you are responsible for having a lawful basis (for example GDPR / ePrivacy in the EU, CAN-SPAM in the US, PECR in the UK) before you contact anyone.
• Marketing rules. Cold e-mail and calls are regulated in many countries. Check the rules for your target country, honour opt-outs, and keep a record of where each contact came from (the release and attribution fields help).
• EU and UK. If you process data of EU or UK residents you are the data controller of your own lead list.
• No guarantees. The data is a snapshot built from several sources, with an estimated confidence per place. Places can have closed or moved.

Data source and attribution

Places come from the Overture Maps Foundation places dataset, read directly from Overture's public cloud storage. The places data is licensed under CDLA-Permissive-2.0; source datasets are listed per row in sourceDataset. If you republish the data, credit it as "© Overture Maps Foundation" and keep the license notice. Learn more at https://docs.overturemaps.org/. City boundaries are looked up with OpenStreetMap Nominatim (© OpenStreetMap contributors, ODbL) and used only to filter places; they are not part of the output.

FAQ

Does the Actor collect usage data? The Actor sends one anonymous usage counter per run (input type, item count range, outcome) to improve the tool; no URLs, content or account data.

Which release is used? The newest Overture release by default (latest); the release name is in every row and in the run summary. You can pin one in the advanced settings. If the newest release cannot be read, the Actor falls back to the last release it was tested with.

Why does my category return nothing? Overture has its own category list. The run log then prints the most common categories in your area; use one of those, or a parent such as food_and_drink. Plain words such as dentist, lawyer, accountant, car repair and their German, French, Spanish, Portuguese, Italian, Dutch, Chinese and Japanese equivalents (Zahnarzt, avocat, abogado, 牙醫, 歯科) are translated for you.

Why did my run end with "None of the categories could be used"? Every category you entered was unreadable (for example only symbols, or a word in a language without a translation). The run ends with one free invalid_input row instead of returning every business in the area. Nothing is charged. Enter the category in English or as an Overture name.

Why are there places from nearby towns? For about 290 major cities (Austin, Berlin, Madrid...) the search covers the urban area, so suburbs such as Round Rock or Potsdam are included. Every row has city, postcode and coordinates, so you can filter them out. Smaller places use their official boundary.

Which phone format do I get? International E.164 (+15122583008) whenever the number is valid for the place's country; otherwise the number exactly as the source has it. phones holds every number of the place, phone the first one.

Why did a location fail with "not a known country"? The part after the comma must be a real ISO country code or an English country name (Berlin, DE, Lisbon, Portugal). UK is accepted and read as GB. A US state is read as the United States (Austin, TX is searched as Austin, US; the state itself is not used and the row note says which place was found). Be careful with state codes that are also country codes (ME is Montenegro, AL Albania): write Portland, US. Several places on one line, separated by ;, are split into separate locations. Failed lines are free.

Why did my box return nothing? A box is written west,south,east,north, so longitude first. Latitude first (59.4,24.7,59.5,24.8) points at the wrong part of the world; the free no_results row then suggests the swapped box when that one has places.

How big can a run be? Cities up to a few hundred thousand places work in one run. Whole countries are possible but slow; always add categories and a Max places limit.

Is this affiliated with any map provider? No. It reads the open Overture dataset only.

Your feedback

Found a wrong or missing category, or want another field? Open an issue on the Actor page.