Skip to main content
Plenoptary
Hardware Flown by DJI M30T + DJI Mavic 4 Pro; H30T thermal: native 1280×1024 radiometric thermal when the M400 flies. · fleet-roster (ADR 0038)Operational discipline scheduled in advance, tracked step-by-step in the portal, and tied to the job state record. · dispatch + value-stream state (ADR 0024/0028)Your file forever You keep the file forever in your portal — every reel, every report, your property's full history. · property-record-portal (ADR 0039)

Features

Reports in hours, with evidence behind every finding.

Every flagged condition shows you the photo, the thermal frame, and the GPS coordinate it came from — so your contractor or carrier can verify the call without re-flying, and so you can argue back when you disagree.

At a glance

Every flagged condition shows you the photo, the thermal frame, and the GPS coordinate it came from — so your contractor or carrier can verify the call without re-flying, and so you can argue back when you disagree.

What the AI actually does

Reads every 100MP photo

A vision model with our prompt frame walks every visual capture and flags conditions a roof and exterior inspection usually cares about: missing or curled shingles, displaced flashing, soffit and fascia damage, gutter sag, debris on the roof. Each finding comes with a short plain-English description and a confidence note — not a verdict.

Reads every thermal frame

A statistical pass over the radiometric thermal frames — real per-pixel temperatures from the DJI Thermal SDK, not a colorized JPEG — flags warm patches that stand out from the rest of the roof. A single hot spot on an otherwise-cool roof gets called out; a uniformly warm south-facing slope does not.

Cross-references the two

Every thermal hotspot back-projects onto the matching 100MP photo using the camera's recorded pose. You see the thermal frame, the wide-RGB photo, and a third image that overlays a marker on the visible roof where the warm patch lives. Three views of the same point, in one row.

What an AI-read report gives you that an eyeball doesn't

A human inspector walks a roof once. The AI reads every frame from every angle the drone captured — and there are hundreds. A roof corner that was photographed from four different orbit points gets four passes, not one. Patterns that only show up across multiple frames (a slow gutter sag, a consistent warm strip along a ridge) surface in the report; one-frame artifacts don't.

Findings also persist. Next year, when you fly the same property, the report compares against this year's findings — same camera positions, same five shots, same prompt frame. You see what changed, not just what's there now.

What the AI can't do, and why that's the honest answer

The AI doesn't make a repair call. It doesn't say "your roof is failing" or "this is covered." It says "here's a warm patch at these coordinates, here's what it looks like in the photo, here's what conditions like this can mean." The judgment about whether to climb, repair, or claim stays with you and your contractor.

That's the design — we'd rather hand you a finding you can argue with than a verdict you have to trust. Every finding shows you the underlying evidence and a confidence note; if the AI flagged something that turns out to be a vent stack doing its job, you can dispute the finding and the report records it.