Local SEO

One word, one country, one button.

A Local SEO campaign finds every business of a given kind in a whole country, reads their website, checks their email actually works, and builds them a listing. Two fields start it. About forty more sit underneath for when you want them — and this page shows you both.

Three ways to start one

They all do the same thing. Use whichever suits the day.

SpiderIQ is an AI-native operating system for agencies — it is built for assistants first. The dashboard exists for people who want to watch it happen with their own eyes. If you would rather just say what you want out loud and have it run, that is the intended way to use it, not a shortcut.

Step one: what, and where

One word in the first box. restaurant. hotel. barber. Not a sentence — the search runs against Google Maps, and a phrase narrows it to nothing.

A country or a US state in the second. The moment you pick one it tells you how many towns it will search. Denmark is 75. A large US state runs to around 950. That number is the honest measure of how big the job is.

Step two: how wide to cast

A whole country is the default, not the only option. Four ways to draw the boundary:

Filter	What it does	Use it when
All locations	Every city in the country	You want total coverage of a market
Cities only	Excludes postcode-level splits	You want breadth without the long tail
By population	Filters by city size	Only the towns big enough to hold your buyer
Specific regions	Pick states, with optional ZIP depth	A territory, a franchise map, one rep's patch

Step three: how deep to read each site

This is the dial that decides both quality and cost, so it is worth understanding. Four modes, each with a real per-site time:

Mode	Depth	Speed	Best for
Quick Contacts	5 pages	~3 sec/site	Cold outreach lists — emails and phones, fast
AI Context	5 pages	~5 sec/site	AI sales agents — builds an AI-ready company document
Lead Research	25 pages	~15 sec/site	Sales research — full intel with team and company
Deep Extract	50 pages	~30 sec/site	Enterprise B2B — maximum coverage of complex sites

And about forty more dials underneath

Every one of them ships already set correctly, which is why two fields are enough to launch. They are there for the campaign where the default is not what you want.

Group	What you can change
Maps	Results per town, search locale across 14 languages, review content, business photos, phone validation, image storage
Crawl	Pages per site, crawl strategy (best-first, breadth-first, depth-first), which page types to prioritise, page timeout
JavaScript sites	Full browser rendering for JavaScript-driven sites, with its own timeout
Extraction	Team members, company information, and AI analysis of business pain points
Summary document	An AI-written markdown summary per business — cleanup level, length cap, and duplicate-section stripping
Lead scoring	Give it your product description and your ideal customer, and every lead comes back scored against them
Email checks	Profile-photo lookup, spam-blacklist checks, emails per business, SMTP timeout
Outreach	Push finished leads straight into an outreach campaign, with per-field mapping

Then five steps run on their own

Step	What it does	Why it takes the time it does
1 · Find	Reads Google Maps for every town in the country	About 14 minutes per town — roughly two thirds of the whole job
2 · Read	Opens each business website and pulls out contacts, team and a description	Runs many at once; this part is quick
3 · Check	Tests every email address before it is used	With step 2, about 8 minutes a batch
4 · Build	Creates a VayaPin directory listing, with the description written in the local language	Only for businesses that cleared step 3
5 · Rescue	No usable email? Finds the Instagram, Facebook or LinkedIn instead	Only runs for the ones step 3 rejected

Note the order. The email is verified before the listing is built, not after — so nothing is created for a business we cannot actually reach.

Where a person still decides

When	The decision
Before	Which kind of business, and which country. This is the whole targeting decision, and it is one word plus one dropdown.
Before	Whether finished leads should be pushed into an outreach campaign. That switch is off by default, and the campaign it points at is locked once the run starts — so leave it off and you will build every listing and email nobody.
After	Read the first twenty results. Are they the kind of business you asked for? Category matching is the one place a Maps search can quietly drift.
After	Answer anyone who replies. Speed here is worth more than anything upstream of it.

Reading the result

The Campaigns tab carries a status and a progress count for every run. Two columns tell you everything.

If it says	It means	Do this
Running	Working. The count climbs as it goes	Nothing. Come back tomorrow
Succeeded	Finished cleanly	Check the count looks sensible for the size of the country
Degraded	Finished, but parts of it struggled	Usable. Worth looking at why
Stopped	Someone halted it	Start a new one if that was not deliberate
Failed	It fell over	Run it again. A re-run of a failed campaign routinely returns a hundred times the results

A campaign that ends Failed is not a dead end and it is not unusual — a country-wide crawl touches a lot of moving parts. The re-run is the fix, and it costs you one click.

Or skip the screen entirely

Every option on this page is also a field your own assistant can set. The same campaign, launched from Python, JavaScript or a single curl:

That is the point of the whole thing. The screen is a convenience; the campaign is an instruction, and instructions are what assistants are good at.