I asked AI to make money for me. It built 51 sites in 10 days.
In August I ran an experiment. I took everything my team learned from building local pay per call sites by hand, wrote it down as rules, and gave it to an AI with access to the Local Sites Pro MCP. This is what happened, what the numbers say ten weeks in, and how you can do the same thing, one chapter at a time. October's figures and a second wave of 17 sites are now included.
By Abhishek Sood · Data: Aug 25 to Oct 31, 2026 · updated Nov 6, 2026 · Published 2026-09-24An AI agent, working through the Local Sites Pro MCP with our written rules, bought and built 51 pay per call websites on expired domains in 10 days. The batch cost $652. By Oct 31, ten weeks after the first site went live, it had taken 464 calls, 38 of them paid, and earned $823, which is 126% of the investment. The first four weeks returned 65%; the batch crossed payback in the last week of October. Individual results vary.
Local Sites Pro is built on our experience from 1,000+ local websites made by me, my team and the LeadsXPro community. LeadSmart reports $614,878 paid out to our community's publishers all-time. You can read the payout report. Individual results vary.
Our team's playbook, replicated by AI.
220 sites
1,452 calls and $5,743 in 90 days. Every call was recorded and transcribed. We learned which services buyers pay for, and wrote down what worked and what did not.
Rules + MCP
No code and no interns. Our lessons written as build rules, access to the Local Sites Pro MCP, and the domains, tracking numbers and Cloudflare account.
51 sites
24 city, 26 nationwide and 1 state site. Median: live 3 days after purchase. Median: first call 5 days after launch. Fastest paid call: 1 day after launch.
Two-thirds back in 30 days. Paid off in nine weeks.
Investment vs payout
- 51 domains, about $10.46 each
- $533
- Local Sites Pro, Pro plan, 1 year
- $119
- Hosting (your own Cloudflare)
- $0
- Tracking numbers and call software (from the network)
- $0
- Buyer payouts, Aug 26 to Sep 24 (month one)
- +$422
- Buyer payouts, Sep 25 to Oct 31
- +$401
Month one returned $422. The next five weeks added $401, and the batch passed its $652 cost in the last week of October, nine weeks after the first site went live. From here, running costs are about $54 a month for all 51 sites, and every dollar above that is profit.
Margin once it is paid back
At October's pace ($315 for the month), 83 cents of each dollar is profit after the $54 running cost. Costs do not grow as calls grow.
- Nationwide sites$560
- City sites$263
Nationwide sites earned 88% of month one, but the city sites caught up: in October they paid $182 against $133 for the nationwide sites, as more of their calls connected to a buyer. 17 of the 51 sites have at least one paid call; 41 have taken a call.
In month one, 142 calls rang and never reached a buyer.
Our city sites got 151 calls in month one, but only 9 connected. The traffic was there. What was missing was a buyer covering those zip codes at that moment. At our network's average conversion rates, those missed calls were worth about $3,300. The fix is to check buyer coverage before you pick the city. October was better: the city sites took 110 calls, 9 of them paid, for $182, and four sites got their first paid call that month (a water heater site in Indiana, a septic site in Illinois, a nationwide roofing site and a commercial refrigeration site).
What the missed calls could have paid
Unconnected calls × network conversion rate × average payout per converted call. An estimate, not money received.
- Plumbing$1,60991 calls
- Water damage$1,14221 calls
- HVAC$37118 calls
- Roofing$1628 calls
- Electrical$554 calls
City sites actually paid $49 in month one. With buyer coverage, the same calls were worth about $3,338.
Use LeadSmart's bid coverage tool.
LeadSmart publishes a coverage page showing the top payout per call for every zip code and vertical, updated daily. In the export I pulled on Aug 30, water damage paid up to $497.50 a call across the Phoenix metro, with Austin at $452, Seattle $442.50, Provo $434 and Miami $398.
- 1
Pick your vertical
Open the coverage tool and choose water damage, plumbing, roofing and so on.
- 2
Sort by payout
Filter by state or city and sort by payout. Note the zips where buyers pay the most.
- 3
Build there
Build your city sites in those places, and write area pages for those exact zips.
- 4
Check again
Payouts change daily, and fewer buyers are live on weekends and outside 8am to 5pm CST.
Estimate what one site could earn.
Pick a vertical and how many calls a month you expect. Conversion rates and payouts are our network averages from 22,106 analysed calls. The top-metro option uses the bid coverage figure above.
Estimates only. Real payouts depend on the zip, time of day and buyer demand.
You do not need to find buyers yourself.
A network gives you tracking numbers, routes your calls to buyers and pays you per qualified call. Join more than one so a call always has somewhere to go.
| Network | What it is | Link |
|---|---|---|
| LeadSmart | Home-services buyers, tracking numbers and a Ringba account with recordings at no charge, plus the zip-level coverage tool. | leadsmartinc.com |
| LeadsMagician | The other network I trust for home-services calls. | Ask in the community |
| eLocal | A long-running local services marketplace with a publisher program. | elocal.com |
| Marketcall | A pay per call network with home-services and other verticals. | marketcall.com |
| OfferVault | A directory. Search pay per call plus your vertical to see offers across many networks. | offervault.com |
“I trust LeadSmart and LeadsMagician. Start with them, and use the others to fill gaps in coverage.” Abhishek Sood
The business earns up to 20× what we get paid for a call.
We transcribe every converted call and pull out the job value the business quotes to the caller. Put that next to what the network pays us per call, and you can see how much room there is for rank and rent.
Median job value to the business
From call transcripts, with what we were paid per converted call.
- Water damage$2,500$121 to us · 21×
- Roofing$850$42 to us · 20×
- Mold$614$45 to us · 14×
- Plumbing$225$30 to us · 8×
- HVAC$125$34 to us · 4×
One water damage call is a median $2,500 job for the restorer, and we get about $121 for it. A restorer who gets three of those calls a month from your site can justify $500 to $1,000 a month in rent.
Start with pay per call, and keep your network in the loop. Selling calls through a network means no buyer hunting, no call-tracking bill and no sales calls while you learn SEO. Once a site's call log proves demand, you have two ways to earn more from it.
Grow it with your network. Tell your network, LeadSmart for example, how many calls you are getting and from which area. They can optimise buyer coverage for that geo, so more of your calls connect and get paid. If you find a local business that wants the calls, you can also bring that buyer to LeadSmart, so the calls keep running through the network and your payout goes up. Ask LeadSmart support how it works for your campaign.
Or rent the site. The same call log is what you show a local business if you prefer a flat monthly fee. See how the pay per call website builder and the rank and rent website builder fit together.
Fresh exact-match domains earn too.
Expired domains with a Google Business Profile behind them are great, but they take practice to find. My team also builds on brand-new exact-match domains. These are the ones that paid out in the last 12 months, with each domain's age so you can plan your own timeline.
Fresh-domain sites that earned
Payout in the last 12 months, with the domain's age today.
- Mold remediation · Texas city$1,05510 months old · 8 paid calls · avg position 24.2
- Water damage · Virginia city$8787 months old · 4 paid calls · avg position 50.6
- Water damage · brandable name$86810 months old · 2 paid calls
- Pipe repair · city site$69410 months old · 11 paid calls · avg position 42.9
- Mold removal · California$6369 months old · 4 paid calls · avg position 6.2
- Water damage · Texas city (.us)$43410 months old · 1 paid call · avg position 46.4
- Water damage · Ohio city$4266 months old · 4 paid calls · avg position 52.5
- Roof moss removal · city site$36710 months old · 9 paid calls · avg position 29
- Basement water removal · state$30210 months old · 1 paid call · avg position 21.5
- Rodent control · city site$28710 months old · 10 paid calls · avg position 45.6
- Water damage · Tennessee city$2516 months old · 1 paid call · avg position 65.7
- Water damage · city site$2396 months old · 1 paid call · avg position 55.2
- Water damage · Arkansas city$2186 months old · 1 paid call · avg position 47.7
- Water damage · Arizona city (.us)$21810 months old · 1 paid call
- Water damage · New York state$9912 months old · 1 paid call · avg position 29.3
- Water damage · Washington town (.us)$5410 months old · 2 paid calls · avg position 21.7
- Water heater repair · city site$5310 months old · 1 paid call · avg position 33.1
Most fresh sites got their first paid call 3 to 6 months after registration.
Four water damage exact-match domains registered on Apr 17, 2026 cost about $44 together. All four have paid out, $1,135 so far, and the first paid call came 84 days after registration.
City + micro-niche keywords.
US monthly searches, cost per click and competition from Google Keyword Planner. Low competition with a high cost per click is where a new site has room.
| water damage restoration miami | Easy | 390 | $53.84 | 0.29 |
| ac repair lake charles | Medium | 390 | $5.85 | 0.58 |
| dumpster rental south bend | Hard | 390 | $1.74 | 0.85 |
| water damage restoration mesa az | Easy | 320 | $9.31 | 0.16 |
| water damage restoration raleigh | Medium | 170 | $53.89 | 0.35 |
| water damage restoration boise | Easy | 170 | $35.78 | 0.34 |
| drain cleaning wichita | Medium | 140 | $7.89 | 0.51 |
| hvac noblesville | Hard | 140 | $37.48 | 0.67 |
| water damage restoration knoxville | Easy | 110 | $26.57 | 0.13 |
| furnace repair madison wi | Easy | 70 | – | 0.13 |
| air conditioning repair springfield il | Hard | 50 | $11.77 | 0.66 |
| water damage restoration peoria il | Easy | 40 | – | 0.09 |
| drain cleaning decatur il | Medium | 10 | – | 0.64 |
| water heater repair florence sc | Easy | 10 | – | 0.14 |
| water damage restoration cedar rapids | Easy | 10 | – | 0.16 |
Easy means competition under 0.35, hard means 0.65 or more. A dash in CPC means Keyword Planner reports no bid data.
Ask your AI which services to build next.
If one niche starts getting calls in a city, build more sites for that same city on different services. Keep only services that are not seasonal, that homeowners cannot do themselves, and that contractors happily take on.
I run pay-per-call lead generation sites. My [NICHE, e.g. plumbing] site in [CITY, STATE] is getting calls, and I want to build more sites in the same city. List 15 specific services within [NICHE] that people in [CITY] search for and call about. For each one tell me: 1. Why demand exists here: housing age, climate, soil, water, local rules, industries. 2. Whether demand is year-round or seasonal. Drop anything mostly seasonal. 3. Whether a homeowner could reasonably do it themselves. Drop anything DIY-able. 4. Whether a licensed [plumber / roofer / HVAC / restoration] company would happily take the job and what it typically costs. 5. A likely search phrase with the city name. Rank the final list by job value, and flag the top 5 to build first.
A local click is worth 100 to 257 times an affiliate click. That is what advertisers pay per click in the US, from Google Keyword Planner. Local businesses pay that much because one customer is worth hundreds or thousands of dollars to them.
Why people are moving into local lead gen.
Affiliate and CPA are getting harder
In my experience, recent Google updates have hit Amazon affiliate and CPA sites hard, and many of those SEOs are looking for their next model.
More pros are moving into local
Experienced SEOs keep entering local lead gen. The cities and micro-niches with no competition today will not stay empty.
Starting costs almost nothing
$119 for a year of unlimited sites is less than one water damage call pays. Waiting is the only expensive part.
Sell the calls, rent the site, or sell the links.
The asset-light start.
The network brings the buyers, the numbers and the call software. You build and rank sites. No buyer hunting, no call-tracking bill, and the recordings show what callers want.
Pay per call websites 2 · Rank and rent · $500 to $1,000 a month is our target rangeRent the site that rings.
Once a city site makes the phone ring, rent it to a local business for a flat monthly fee. Your call log is the sales pitch, and one tenant covers the costs of dozens of sites.
Rank and rent websites 3 · Service-niche blogs · $20 to $100 per link is a typical rangeSell niche-relevant links.
Local Sites Pro lets you add blog posts. A real-looking service site ranks for low-competition service questions, and an aged, topical site becomes a place to sell links and support your own money sites.
Watch: service niche bloggingBuild 30 sites. See what they could make.
Move the sliders. The defaults use our real data where we have it and clearly marked assumptions where we do not.
Costs: $11 per domain per year and $119 a year for the Local Sites Pro Pro plan (unlimited sites). Hosting on your own Cloudflare is free. Pay per call income ramps up over the first 3 months, rent ramps from month 4 to 6, and links start in month 3. These are scenarios, not guarantees: rankings, buyer coverage and tenants vary by niche and city.
Where ranked sites get sold.
Flippa
List single sites or a whole network with call logs and payout history.
Facebook groups
Rank and rent and local SEO groups buy proven city sites.
Local agencies
Agencies buy ranking sites to place their own clients.
Proof it works
We sold a fresh-domain local lead gen site for $5,500.
Watch the case studyBought, live, first call, first paid call.
Every site in the batch on one timeline, through Oct 31. Domains are partly masked.
Site-by-site data, through Oct 31.
Calls, paid calls and payout are cumulative from launch to Oct 31, 2026. The October columns show that month alone. Indexed pages are from the Oct 25 check. Select a column heading to sort.
| rep••••ng.com | City | Roofing · Parma OH | Aug 25 | Aug 27 | Aug 27 | Sep 14 | 1 | 19 | 4 | $127.83 | 4 | $63.57 |
| aaa••••mi.org | Nationwide | Electrical repair | Aug 25 | Sep 1 | Sep 10 | Sep 10 | 2,840 | 14 | 4 | $111.99 | 4 | $22.68 |
| mel••••ng.co | Nationwide | Water heater repair | – | Aug 29 | Sep 10 | Sep 10 | 1,970 | 4 | 2 | $103.95 | 2 | $50.40 |
| cle••••es.com | Nationwide | Porta potty + dumpster | Aug 25 | Aug 27 | Aug 28 | Aug 28 | 9,260 | 17 | 8 | $68.00 | 4 | $8.50 |
| sha••••ic.com | Nationwide | Generator install | Aug 25 | Aug 30 | Sep 15 | Sep 15 | 10,900 | 12 | 3 | $57.96 | 5 | $29.61 |
| jkh••••ac.com | Nationwide | Air duct cleaning | Aug 25 | Aug 30 | Sep 3 | Sep 23 | 1,840 | 11 | 1 | $47.25 | 0 | – |
| tha••••nc.com | Nationwide | HVAC | Aug 25 | Aug 27 | Aug 30 | Aug 30 | 7,800 | 3 | 1 | $45.90 | 1 | – |
| thi••••ng.com | Nationwide | Plumbing | Aug 25 | Aug 27 | Aug 28 | – | 12,500 | 16 | 1 | $45.50 | 9 | – |
| and••••nc.com | City | HVAC · Madison WI | Aug 25 | Aug 27 | Aug 28 | Sep 5 | 58 | 32 | 4 | $37.96 | 21 | $33.77 |
| aft••••lc.com | City | Drain cleaning · Duluth MN | Aug 25 | Aug 26 | Aug 26 | Sep 4 | 1 | 36 | 3 | $37.00 | 16 | $24.40 |
| wat••••ce.com | City | Water heater · Logansport IN | Aug 25 | Aug 27 | Sep 3 | Oct 22 | 38 | 8 | 1 | $31.50 | 1 | $31.50 |
| kel••••ce.com | City | Septic/sewer · Decatur IL | Aug 25 | Aug 27 | Aug 27 | Oct 5 | 116 | 40 | 1 | $28.35 | 22 | $28.35 |
| gtr••••lc.com | Nationwide | Flat roofing | Aug 25 | Aug 30 | Sep 19 | Sep 19 | 1,140 | 2 | 1 | $22.57 | 0 | – |
| all••••lc.com | Nationwide | Metal roofing | Aug 25 | Aug 26 | Aug 27 | Sep 1 | 14,600 | 18 | 1 | $18.90 | 1 | – |
| fle••••co.com | Nationwide | Mold remediation | Aug 30 | Aug 31 | Sep 10 | Sep 21 | 9,670 | 20 | 1 | $16.36 | 10 | – |
| sup••••rs.com | Nationwide | Roof replacement / siding | Aug 25 | Aug 29 | Oct 2 | Oct 19 | 14,900 | 8 | 1 | $15.75 | 8 | $15.75 |
| jrc••••ng.com | Nationwide | Commercial refrigeration | Aug 25 | Sep 1 | Sep 7 | Oct 5 | 4,550 | 16 | 1 | $6.05 | 5 | $6.05 |
| aff••••in.com | City | HVAC · Noblesville IN | Aug 25 | Aug 26 | Sep 2 | – | 82 | 9 | 0 | – | 4 | – |
| am-••••lc.com | City | HVAC · Lake Charles LA | Aug 25 | Aug 26 | Aug 28 | – | 121 | 5 | 0 | – | 0 | – |
| gro••••ic.com | Nationwide | Septic service | Aug 25 | Aug 26 | Sep 5 | – | 80 | 4 | 0 | – | 0 | – |
| cha••••ng.com | City | Septic · Rio Rancho NM | Aug 25 | Aug 27 | Sep 2 | – | 10 | 18 | 0 | – | 7 | – |
| dis••••es.com | City | Plumbing · Orange City FL | Aug 25 | Aug 27 | Aug 28 | – | 25 | 10 | 0 | – | 3 | – |
| drw••••on.com | City | Water damage · Cedar Rapids IA | Aug 25 | Aug 27 | Sep 3 | – | 1 | 6 | 0 | – | 1 | – |
| ele••••os.us | Nationwide | Electrician | Aug 25 | Aug 27 | Sep 15 | – | 3,990 | 5 | 0 | – | 0 | – |
| lat••••er.com | Nationwide | Emergency roof repair | Aug 25 | Aug 27 | – | – | 8,610 | 0 | 0 | – | 0 | – |
| sup••••ol.com | Nationwide | HVAC | Aug 25 | Aug 27 | Aug 29 | – | 305 | 12 | 0 | – | 6 | – |
| fel••••ng.com | City | Plumbing · Norristown PA | Aug 25 | Aug 28 | Sep 3 | – | 30 | 13 | 0 | – | 6 | – |
| hun••••s2.com | City | Plumbing · Florence SC | Aug 25 | Aug 28 | Sep 4 | – | 29 | 2 | 0 | – | 0 | – |
| jpl••••gg.com | City | Plumbing · Binghamton NY | Aug 25 | Aug 28 | Sep 3 | – | 40 | 10 | 0 | – | 5 | – |
| bar••••mo.com | City | Plumbing · Poplar Bluff MO | Aug 25 | Aug 29 | Sep 3 | – | 26 | 15 | 0 | – | 10 | – |
| dra••••ky.com | Nationwide | Garbage disposal repair | Aug 25 | Aug 29 | – | – | 7,150 | 0 | 0 | – | 0 | – |
| eff••••er.com | Nationwide | Sump / well pump | Aug 25 | Aug 29 | Sep 5 | – | 749 | 1 | 0 | – | 0 | – |
| pat••••ng.com | City | Plumbing · Guthrie OK | Aug 25 | Aug 29 | Sep 3 | – | 27 | 9 | 0 | – | 3 | – |
| pre••••nc.com | Nationwide | Water damage restoration | Aug 25 | Aug 29 | – | – | 394 | 0 | 0 | – | 0 | – |
| tri••••si.com | City | Plumbing · Vevay IN | Aug 25 | Aug 29 | Sep 1 | – | 10 | 12 | 0 | – | 0 | – |
| cro••••ir.com | Nationwide | Heat pump install | Aug 25 | Aug 30 | – | – | 1,230 | 0 | 0 | – | 0 | – |
| dum••••ta.com | Nationwide | Roll-off dumpster | Aug 25 | Aug 30 | – | – | 8,160 | 0 | 0 | – | 0 | – |
| gre••••an.com | Nationwide | Water damage restoration | Aug 30 | Aug 31 | Sep 4 | – | 12,200 | 2 | 0 | – | 0 | – |
| mon••••lc.com | Nationwide | Sewage cleanup | Aug 30 | Aug 31 | Sep 2 | – | 10,400 | 11 | 0 | – | 1 | – |
| cen••••rk.com | City | HVAC · Springfield IL | Aug 25 | Sep 1 | Sep 3 | – | 55 | 4 | 0 | – | 0 | – |
| car••••ls.com | State | Roll-off dumpster (NC) | Sep 1 | Sep 1 | – | – | 496 | 0 | 0 | – | 0 | – |
| dum••••ub.com | City | Dumpster · South Bend IN | Sep 1 | Sep 1 | – | – | 58 | 0 | 0 | – | 0 | – |
| joe••••ng.com | City | Water damage · Casa Grande AZ | Aug 30 | Sep 2 | Sep 3 | – | 37 | 9 | 0 | – | 3 | – |
| spo••••ce.com | City | Water damage · Mesa AZ | Aug 30 | Sep 2 | Sep 3 | – | 57 | 6 | 0 | – | 1 | – |
| sec••••fl.com | City | Electrician · Ocala FL | Aug 25 | Sep 3 | Sep 10 | – | 39 | 7 | 0 | – | 3 | – |
| car••••us.com | City | Water damage · Miami FL | Aug 30 | Sep 3 | Sep 10 | – | 28 | 5 | 0 | – | 0 | – |
| dre••••tx.com | Nationwide | Fire damage restoration | Aug 30 | Sep 3 | – | – | 7,280 | 0 | 0 | – | 0 | – |
| mil••••ng.com | City | Water damage · Raleigh NC | Aug 30 | Sep 3 | – | – | 123 | 0 | 0 | – | 0 | – |
| dra••••ta.com | City | Plumbing · Wichita KS | Aug 25 | Sep 4 | Sep 9 | – | 97 | 5 | 0 | – | 0 | – |
| atx••••rs.com | Nationwide | LVP flooring install | Aug 30 | Sep 4 | – | – | 1 | 0 | 0 | – | 0 | – |
| myc••••em.net | Nationwide | Carpet installation | Aug 30 | Sep 4 | Sep 21 | – | 2,130 | 8 | 0 | – | 1 | – |
The AI kept building. Ten more sites went live on Sep 25 (6 city, 4 nationwide) and seven nationwide sites on Nov 2 to 5, all through the same MCP and the same rules, on the same Pro plan. The September ten took 41 calls by Oct 31, and a nationwide HVAC site got its first paid call 9 days after launch. The November seven are too new to have data. Individual results vary.
Second wave, site by site.
Launched Sep 25 and Nov 2 to 5, 2026. Calls, paid calls and payout are from launch to Oct 31. Domains are partly masked.
| mrm••••es.org | Nationwide | Mosquito and pest control | Nov 5 | – | – | – | 0 | 0 | – |
| pro••••ol.com | Nationwide | Pest control | Nov 5 | – | – | – | 0 | 0 | – |
| a1h••••es.com | Nationwide | Plumbing | Nov 2 | – | – | – | 0 | 0 | – |
| bea••••ty.com | Nationwide | Plumbing | Nov 2 | – | – | – | 0 | 0 | – |
| rel••••ut.com | Nationwide | HVAC | Nov 2 | – | – | – | 0 | 0 | – |
| spb••••ro.com | Nationwide | Plumbing | Nov 2 | – | – | – | 0 | 0 | – |
| the••••an.com | Nationwide | Plumbing | Nov 2 | – | – | – | 0 | 0 | – |
| aff••••ed.com | Nationwide | HVAC | Sep 25 | Oct 4 | Oct 4 | 7 | 6 | 1 | $44.23 |
| air••••fl.com | City | AC repair · Port Charlotte FL | Sep 25 | – | – | 2 | 0 | 0 | – |
| ant••••ng.com | City | Water damage · Illinois city | Sep 25 | – | – | 29 | 0 | 0 | – |
| cer••••mi.com | Nationwide | Commercial roof coating | Sep 25 | – | – | 7,120 | 13 | 0 | – |
| eli••••ng.net | Nationwide | Laminate flooring install | Sep 25 | – | – | 1,180 | 1 | 0 | – |
| jrs••••ng.com | City | Water damage · Illinois city | Sep 25 | – | – | 38 | 1 | 0 | – |
| mos••••ng.com | City | Roofing · Idaho city | Sep 25 | – | – | 36 | 12 | 0 | – |
| str••••ng.com | City | Water damage · Utah city | Sep 25 | – | – | 38 | 2 | 0 | – |
| the••••ks.org | Nationwide | Roofing | Sep 25 | – | – | 5,250 | 2 | 0 | – |
| war••••re.com | City | Water damage · Utah city | Sep 25 | – | – | 76 | 4 | 0 | – |
First-call dates for the second wave are shown where we pulled the call log; a blank means not checked, not zero calls.
Do it yourself with the Local Sites Pro MCP.
Connect the Local Sites Pro MCP to Claude or ChatGPT, give it your rules, and work through these steps in order.
- 1
Connect everything once
Set up Local Sites Pro with your Cloudflare account. Connect your Google Search Console and Ringba accounts, so your sites, search data and calls all show in one dashboard and your AI can read them over MCP.
get_setup_statusgsc_status - 2
Write your rules
Narrow services that buyers pay for, one buyer intent per site, no fake reviews, no keyword cannibalisation, and a different look for every site. Listen to your call recordings. The services callers ask for are your micro-niches.
gsc_performancegsc_opportunities - 3
Pick the city by buyer coverage
Check the bid coverage tool for your vertical and choose zips where buyers pay well right now.
bid coverage - 4
Find a domain
Search expireddomains.net for niche keywords and filter for domains that were a local business. Check them in Semrush: a local domain that still ranks for keywords usually had a Google Business Profile. No luck? A fresh exact-match domain works too.
expireddomains.netSemrush - 5
Get the tracking number, then check the domain
Get a number from your network: local for a city site, toll-free for nationwide. Make sure the domain does not already have a live site. Rule: no tracking number, no build.
list_websitesget_website_status - 6
Build in manual mode, and let Claude or ChatGPT write
Claude and ChatGPT write better local content than any template prompt. Create the site in manual content mode through the MCP, then have your AI write every page itself for that exact niche and city. The more context you give, the better it gets. Nationwide sites with hundreds of city pages are available to LeadsXPro members.
create_websitecreate_nationwide_website - 7
Polish, link and publish
Add service pages and blog posts, give each site its own design, link sister sites together, and publish.
update_page_contentadd_website_contentupdate_website_designpublish_website - 8
Measure, then double down
Watch indexing and calls weekly. At day 45 and day 90, rebuild what stalls, rent out what rings, and build more sites in the cities that work.
gsc_inspect_urlgsc_submit_sitemaprewrite_website_pages
The brief to give your AI.
Paste this into Claude or ChatGPT with the Local Sites Pro MCP connected. Fill in the brackets.
Your sites are yours. The Pro plan lets you export any website as static HTML, so you can host it anywhere or hand it over when you sell it.
You are writing a local service website through the Local Sites Pro MCP. Create the site in manual content mode and write every page yourself. Don't leave anything to a generator. THE BUSINESS - Niche: [e.g. water damage restoration] - Main service (primary keyword): [e.g. water damage restoration] - Second service: [e.g. water extraction] - City and state: [CITY, ST] · service area: [nearby towns and neighbourhoods] - Tracking phone: [NUMBER] · company name: [NAME, never a person's name] WHAT TO RESEARCH AND USE - Local specifics: housing age and types, climate and storms, soil, water hardness, local codes or permits, common problems in this city. - Real neighbourhoods, landmarks and zip codes in the service area. - What callers actually ask: [paste phrases from your call recordings]. - The services a licensed [contractor type] here takes on, and rough price ranges. PAGES - Home, 10-15 service pages (one narrow service each), 10+ area pages, 6-10 FAQs, 4-6 blog posts on long-tail questions. - Every service page: what the problem looks like, what the job involves, what affects the price, when it's an emergency, and one local detail only this city would have. - Link between service pages inside the text, not just in menus. RULES - Company voice. No owner names, no reviews or ratings, no invented licenses, awards or guarantees. - Don't target the same keyword + city on two pages. - Write like a local tradesperson talking to a homeowner: plain words, varied sentence length, no filler, no em dashes. - Titles: [Service] in [City], [ST] | [Company]. Meta descriptions under 155 characters with the city and a reason to call. When the site is built, give it its own colours and font, publish it, and show me the live URL.
How to find the domains, on video.
Find expired domains for pay per call, rank and rent and local business
Where to look and what to filter for.
Watch Short · Jul 2026How I validate an expired domain before building
The checks I run before I buy.
Watch Walkthrough · full tutorialExpiredDomains.net filters for niche-specific indexed domains
The exact filter setup.
Watch Short · quick tipFind expired domains with keywords and traffic using Semrush data
Using keyword and traffic data to shortlist.
Watch Case study · May 2025$6,500+ in under 3 months with fresh domains, then sold for $5,500
A fresh-domain site from build to sale. Individual results vary.
Watch Playlist · YouTubeThe full Local Sites Pro course
Every tutorial in one playlist, at no charge.
Open the playlist12 · What it costs: less than one water damage call.
The 51 sites in this study ran on one Pro plan. Hosting stays free on your own Cloudflare.
- 3 local websites (lifetime)
- Single-city & single-state sites
- AI content, images & styles
- Free hosting on your Cloudflare
- Unlimited local websites
- Search Console + AI optimization
- Full prompt library · static HTML export
- Priority support
Five things to do before you build ten sites.
- 1
Join a network
Join a pay per call network and check its bid coverage for your vertical.
- 2
Pick one city, one service
Choose one high-payout city and one narrow service.
- 3
Find 3 to 5 domains
Expired with a Google Business Profile history, or fresh exact-match.
- 4
Get numbers, then build
Get a tracking number for each, then build in manual mode with Claude or ChatGPT writing.
- 5
Publish and check
Publish, link the sites together, and check indexing next week.
Where the numbers come from
- The 51 sites: built by an AI agent through the Local Sites Pro MCP on expired domains of former local businesses, Aug 26 to Sep 4, 2026. Purchase dates come from the domain registry, calls and payouts from our call tracking (Ringba through the network) from launch to Oct 31, 2026, and indexed pages from Local Sites Pro's weekly Google check on Oct 25, 2026. The first version of this page covered Aug 26 to Sep 24 (255 calls, 19 paid, $422); this update adds Sep 25 to Oct 31.
- Indexed pages: 166k at the Sep 24 check, 156k on Oct 25. Google trimmed some of the nationwide sites' thinner city pages over October while the city sites went from 875 to 1,109 indexed pages. Nationwide sites: 154,649 indexed, median 7,150 per site.
- The second wave: 17 more sites on the same account, built the same way: 10 live on Sep 25 and 7 on Nov 2 to 5, 2026. Calls and payouts through Oct 31.
- Team baseline: 220 sites built by our interns, last 90 days.
- Missed-call value: 142 unconnected calls × our network's average conversion rate and payout per niche. An estimate of what coverage could have paid, not money received.
- Job values: median job value quoted in our transcribed converted calls, next to our average payout per converted call in the same niche.
- Fresh-domain sites: our team's sites on new exact-match domains, payout over the last 12 months. Ages come from registry registration dates.
- Bid coverage figures: from a LeadSmart coverage export on Aug 30, 2026. Payouts change daily.
- Keywords: US volume, CPC and competition from Google Keyword Planner via Keywords Everywhere, Sep 2026. Positions are Search Console 90-day averages.
- Calculator: scenarios from the figures above plus stated assumptions. Past results do not guarantee future earnings. Domains are partly masked.
- Nationwide sites: the nationwide engine is available to LeadsXPro members. Starter and Pro build city and state sites.
For the wider dataset, including sites built before this batch, see the 72-site pay per call case study.
Questions about this experiment.
Can AI really build websites that make money?
In this test, yes, at a small scale. An AI agent built 51 sites through the Local Sites Pro MCP and the batch earned $823 in its first ten weeks against $652 invested, passing payback in week nine. Individual results vary.
What did the AI actually do?
It followed our written rules, created each site through the MCP, wrote the pages, gave each site its own design and published it. People chose the domains, the networks and the rules.
How much did it cost to build 51 sites?
$652: about $533 for 51 domains and $119 for one year of the Pro plan. Hosting was $0 on Cloudflare and the network supplied the tracking numbers.
How long until the first call?
The median site got its first call 5 days after launch, and the fastest paid call came 1 day after launch. Some sites took longer: four of the 51 got their first paid call in October, five to eight weeks after launch.
Did the money keep coming after month one?
Yes. Month one paid $422. Sep 25 to Oct 31 paid another $401, with October alone at $315 from 167 calls. 17 of the 51 sites have at least one paid call so far.
Which AI assistant do I need?
Claude works with the MCP server out of the box. ChatGPT needs Developer Mode enabled.
Will I get the same results?
No one can promise that. Results depend on the niche, the city, the domain and buyer coverage. Individual results vary.
Give your AI the brief. Your first site is ten minutes away.
Connect the MCP serverKeep going from here.
- Pay per call affiliate marketing with local sitesThe money page: how the sites, numbers and networks fit together.
- Rank and rent websitesThe builder made for ranking a site, then renting it.
- Lead generation websites for local lead generationBuild a local lead gen business on sites you own.
- Best pay per call affiliate networksLeadSmart, Leads Magician, eLocal and the rest, compared.
- What is rank and rent?The business model, the math and the first niche.
- How to start a local lead generation businessFrom the first site to a network.