How LeadForge works, step by step.

Choose the country and place

Start with the country selector, then type a city or region name. Suggestions come from a city index of roughly 70,000 places, built from GeoNames and shipped inside the application, so the list responds instantly and your keystrokes are not sent to an external geocoder. Choosing a suggestion sets both the place name and the coordinates that anchor the search — the application never guesses a location from free text.

The search opens on Brisbane, Queensland by default so there is always something to run. If you would rather start from where you are, Use my current location asks the browser for permission, then matches those coordinates to the nearest city in the same local index.

A city centroid is rarely the exact centre of the area you care about. The map beside the form is interactive: click anywhere or drag the pin to move the search centre, and the coordinates above the map update as you go. This matters more than it looks — a 10 km circle drawn from a city centre and a 10 km circle drawn from a suburb cover very different sets of businesses.

Choose the business types

You can select up to three types per search. The built-in list covers 25 common categories — dentist, restaurant, café, pharmacy, gym, accountant, lawyer, real estate agent, car repair, hotel, barber, beauty salon, supermarket, bakery, veterinarian, school, clinic, hospital, IT company, marketing agency, bank, fuel station, bus station, electronics store and physiotherapist — and the filter box above them searches both the labels and their aliases, so typing chemist finds the pharmacy category and law firm finds lawyers.

Each category maps to specific tags in the underlying map data. That is why the categories behave predictably: choosing Pharmacy is not a keyword search for the word pharmacy, it is a request for features tagged as pharmacies.

When nothing in the list fits, the custom type field takes any term between 2 and 60 characters — a type such as bus station, or a specific business name. A custom term is matched two ways at once: against feature names, and against the common tagging keys for amenities, shops, offices, tourism, leisure, healthcare and public transport. That dual approach is what lets a single box handle both “a kind of place” and “this particular chain”.

The counter at the bottom of the panel keeps the three selections, the radius and the result cap visible in one line before you commit to a search.

Set the radius and the result cap

The radius slider runs from 1 to 50 kilometres and the result cap from 10 to 200 businesses. Both change the cost of the query, not just its size. A small radius returns faster and gives you a tighter, more workable list; a 50 km radius over a dense city can take noticeably longer and may return more records than you will realistically review.

A useful habit is to start at 5–10 km. If the area turns out to be thinly mapped, widen it; if you are drowning in results, narrow it or drop a category rather than raising the cap.

What happens when you press search

The application sends your parameters to its own API, which builds a geographically bounded query — one clause per category, each restricted to your radius around your coordinates — and sends it to a public Overpass endpoint, the standard query service for OpenStreetMap data. Nothing about you goes with it; the query describes an area and a set of tags.

Three things then protect the shared service everyone depends on. Identical queries are answered from a short-lived server-side cache rather than repeated upstream. Searches are rate limited per visitor. And the query itself carries a timeout, so a request that cannot be answered fails cleanly instead of hanging. If you see a message asking you to wait, or suggesting a smaller radius, that is one of these limits doing its job — the public map service is a community resource, not a commercial API with capacity to burn.

Typical searches complete in a handful of seconds. Large radii with several categories take longer, and a search that exceeds the client-side time limit will tell you so and suggest a 5–10 km radius.

Reading the results

Each raw map feature is converted into a business record before you see it. The name comes from the feature name, or a brand or operator when no name is set. The address is assembled from individual address components — house number, street, suburb, city, state, postcode — or taken whole where a full address is recorded. Phone numbers, websites, e-mail addresses, social links and opening hours are pulled from their standard tags, and the straight-line distance from your search centre is calculated for every record.

Before the list reaches you it is ordered so that records with a phone number or website come first and, within that, the nearest come first — then it is cut to your result cap. That ordering is deliberate: if you asked for 50 businesses out of 300 candidates, the 50 you get are the ones you can actually act on.

Duplicates are merged on the way through. Two records are treated as the same business if they share a source identifier, or share a name and address, or share a name and website domain. When records merge, populated fields win over empty ones, so the surviving record is the more complete of the two.

Every card shows a completeness percentage. It is a plain arithmetic description of the record, not a judgement about the business: a listing starts at 20 and gains 15 for an address, 20 for a phone number, 20 for a website, 15 for an e-mail address, 5 for opening hours and 5 for coordinates. A 100% business is not a better prospect than a 40% one — it is simply a record you have to do less work to verify.

The four counters above the list — total businesses, phone numbers, websites, e-mails — recount as you filter, which makes them a quick way to test whether an area is worth pursuing at all before you start reading individual cards.

List, map, or both

The same result set can be displayed three ways: as a list, as a full-width map, or split between the two. The views are linked — selecting a card focuses the corresponding point on the map, and clicking a point on the map highlights its card. The split view is the one to use when geography matters: clusters, gaps and the businesses that sit just outside the area you actually service become obvious in a way a list never shows.

Above the list, filters narrow what both views display: free-text search across name, address, category and city; a category filter; a city filter; a distance filter at 5, 10, 25 or 50 km; and presence toggles for phone, e-mail, website, social profiles and opening hours. Sorting offers nearest first, business name, or most complete. Filtering changes what is shown, what the counters report, and what the export buttons write out — which is exactly how you produce a targeted file without editing anything by hand.

Saving a shortlist

Select the businesses worth keeping and use Save to store them in this browser. Saved records go to IndexedDB — up to 200 of them — and appear under the Saved view, where the same filters, sorting, contact discovery and exports are available. The last five searches are kept too, with their results.

There is no account, so this is genuinely local: the list does not follow you to another browser or device, nobody else can see it, and clearing your browser data or using Clear my local data on the About page removes it permanently. Export anything you would be annoyed to lose.

You can also add a business by hand — useful for a referral or a company that simply is not mapped. It is stored alongside the rest and exports the same way, tagged so you can tell manually entered records from map records later.

Optional contact discovery

Many listings have a website but no e-mail address. Contact discovery closes that gap on demand: select up to ten records and use Find contacts, or run it on a single card. The button is disabled when a listing has no website, because there is nowhere to look.

For each selected business the server visits that business’s own public homepage and, if one is linked, at most one same-site contact or about page. It reads what is publicly visible — e-mail addresses, telephone links, links to social profiles — and merges them into the record. It fetches only public web addresses, refuses private network destinations, follows only a small number of redirects, stops at a size limit, and does not attempt to get past logins or other access controls. A few requests run at a time, not hundreds.

It will not always succeed. Sites that are down, that block automated requests, or that publish contact details only in an image or a form will come back with nothing found — which is a result, not an error.

Verify before you use anything

This is the step people skip. Open the listing’s website and confirm the business still trades under that name at that address. Check that the phone number on the site matches the one in the record. Confirm that an address found by contact discovery is the right one to write to — a generic inbox is usually a better destination than a named person’s address scraped from a team page. Every card links to the underlying map record so you can see the source directly.

Verification is also the moment to drop records. A list of 40 confirmed businesses is worth more than 200 unverified rows, and deleting the ones that do not fit is far cheaper before an export than after.

Export what you selected

Exports are generated in your browser and download immediately. You can write the whole filtered list, only the records you selected, only those with phone numbers, or only those with e-mail addresses — plus GeoJSON when you want geographic features rather than rows. The CSV carries the source and source identifier for every record, so a row can always be traced back to where it came from. Exporting business data covers the columns and the clean-up worth doing first.

Know what the tool cannot tell you

LeadForge reports what is mapped. It does not know whether a business is trading, whether the phone number still reaches someone, or whether anyone there wants to hear from you. Coverage varies enormously between regions and categories, records can be years out of date, and an empty result set often means “nobody has mapped this yet” rather than “nothing is here”.

Read Working with open business data to understand why the gaps exist, and Responsible business outreach before you contact anyone on the list.