Where should the site go? NOT BUILT YET

Coverage answers a question about a site you can already name. The site planner answers the earlier question: you know the area that needs service and the band you have, but not yet which structure you will mount on.

What it does

1. Finds candidate structures. Registered towers in and around your area, from the FCC Antenna Structure Registration database.
2. Works out what height is free. Existing licensed antennas on each structure are read from the Universal Licensing System, so the mount bands nobody is using can be identified.
3. Renders every candidate. Each one gets a coarse coverage prediction on the same engine as the paid product, at a resolution chosen for comparison rather than for delivery.
4. Picks the smallest set that covers the area. Rather than the best single site, it selects the fewest sites that together provide the coverage, which is usually a different answer.

What you will be able to ask for

What comes out

• A coverage map of the selected set, plus one per site showing what that site alone contributes.
• A KMZ with the map as a ground overlay and a placemark on each chosen site — structure type and height, the mount band that looked free, existing tenants by service, the owner's contact details, and what to ask them for.

The contact list is included in the KMZ rather than provided as a separate download. A single file carries the map and the reasoning together, opens in common mapping applications, and avoids producing a bulk contact list: each contact is attached to a place on the map, which is how it will be used.

Ask it in plain language

The planner is built as a set of tools an AI assistant can use, so the conversation is the interface: “two VHF sites for the county, handheld outdoors”, then “without the Allen tower”, then “show me the map”. Follow-up questions cost nothing once an area has been scouted. Your assistant can view the maps and can be given photos. Connecting your own assistant →

What it will not tell you

These limitations will remain when the planner is finished. They are properties of the public data it relies on.

• Tenancy is evidence of tenants, not of space. Knowing what is licensed on a tower does not tell you the owner will rent to you, what it costs, or whether the structure can carry more.
• The registry is not the world. Cellular is licensed by market rather than by site and never appears. Unregistered structures — most buildings, most short masts — are not in it at all.
• Joining licences to structures is approximate. Many older public-safety grants carry a free-text location and no structure number, so matching is done by coordinate. On one measured structure, matching by coordinate found 18 licensed antennas where matching by registry number found 9. Relying on the registry number alone would have understated that tower's tenancy by half.
• The scouting renders are deliberately coarse. They exist to rank candidates against each other, not to be delivered. The chosen sites are worth re-rendering properly afterwards, and that is a paid render.
• "More coverage over the hospital" has a price. Weighting one place above the rest can change which sites are chosen. The answer always states what that costs elsewhere, and states clearly when no candidate structure can reach the place at all. The percentages it reports are coverage, not weighted demand.
• A mast you invent is not a registry structure. None of the ownership, tenancy or availability language applies to a point you placed yourself. It is a what-if about terrain.

Related reading

• Coverage — the propagation engine underneath, and what it models.
• Resolution & cost — why the scouting renders are coarse on purpose.
• Pricing — how a finished render is charged.