Whether it's five photos or five hundred, most of the time isn't spent on engineering judgement. It's spent on file management, photo numbering, code lookup, and Word formatting — work that adds nothing to the deliverable, regardless of how big the job is.
IMG_4837.jpg to 03-north-ladder.jpg — it just gets doneTAR isn't a "big project" tool — it's a workflow. Setup overhead is zero, so it's worth using on the five-photo retainer check, and it scales without effort to the five-thousand-photo portfolio survey.
Same drag-and-drop. Same keyboard. Same firm-branded PDF. No matter the size.
Every kind of inspection has its own language. TAR carries a tag set per project — pick the right one when you create the project and the codes for this inspection light up on the keyboard.
1,847 iPhone photos dropped onto TAR. In ~90 seconds: clustered into 38 GPS-bounded areas, each with a centroid for the cover map, an AI-drafted caption per photo, and visible-text OCR (panel labels, addresses, drawing numbers) captured. You then tag with one keystroke per finding instead of building the report scaffold by hand.
Hybrid clustering produces 12 areas. You re-name 4, merge 2, split 1 at a gap badge — done in 8 minutes. Same scaffold, scaled down. No setup overhead.
All four preserve manually-grouped areas. Re-cluster as many times as you want — your hand-grouping is never destroyed.
Point at the photo folder. TAR reads images in place — no copy, no upload. The project remembers where you left off the next time you open it.
Auto-cluster runs on import. Drag thumbnails between areas, click a gap badge to split, click between areas to merge. Propagate Names copies area labels down to photo titles in one click.
Press a letter, draw a box. Codes, severities, regulation references attach automatically. Shift+letter for a whole-image tag. Hands never leave the keyboard.
Pick the template (PDF or Word). Click export. Image numbers, captions, map page, tag legend — all generated, all numbered, all in firm style.
No mode-switching paperwork. Move between Group, Tag, and Report from a single header tab.
Photos with GPS auto-pin on an OpenStreetMap basemap rendered locally — no API key, no cloud, no tracking. Pins use the firm colour and stack when locations overlap so the map stays readable on a 500-photo inspection. Tile fetching falls back to a clean schematic if you're offline.
The map page goes straight into the report as a navigation overview. Pin numbers cross-reference the image pages so the reviewer can jump from a finding on the map to the photo that documents it.
320 photos pinned on the basemap; the cover map shows 14 area centroids with colour-coded severity, cross-referenced by pin number to image pages.
Pins trace your walking path along the deck and underside; the basemap proves coverage to the client without a separate site-walk diagram.
Photos without GPS — interiors, basements, roof voids, mech rooms — get pinned on whatever plan you load: a PDF floorplan, a scanned sketch, even a hand-drawn schematic. Click on the plan to drop a pin for the active photo.
Same pin style as the GPS basemap, same cross-references into the report. One consistent visual language across outdoor and indoor sections of the same deliverable.
220 interior photos pinned on per-floor PDF plans. The report renders one plan per area with the area's photos numbered around it — reviewers walk the building on paper.
Hand-drawn schematic loaded; pins mark each access hatch and noted defect. No GPS, no problem — the report still has a navigable visual index.
A vision model drafts a caption for each photo on import — what's in it, the apparent condition, anything notable. You edit instead of drafting.
OCR reads visible text out of every photo: serial numbers, panel labels, address signs, equipment tags, drawing numbers. Captured automatically.
Find photos by what's in them, not just by tag. "Panel 3B", "rusted railing", "exposed conduit" — query in plain English, jump to the matching photo.
A local model flags likely objects of interest — ladders, panels, structural members — and proposes the matching tag for you to confirm.
Every project carries its own context, so AI suggestions and the report itself are scoped to this inspection — not your whole laptop.
Per-project tag set — a different vocabulary per inspection type, switched per project
Per-project branding — firm name, colours, logo, organization carry through to the report cover
Multi-user awareness — others can't stomp your write while you're editing the same project
Resume where you left off — captions, titles, areas all scoped to the project; reopen and pick up exactly there
AI features run locally on your laptop. No upload, no cloud, no privilege concerns. Photos and analysis stay on the same machine.
Point at a folder. TAR reads it in place. No copy step.
Time, GPS, hybrid, or folder. Hundreds of photos → named areas.
Press a letter, draw a box. Codes, severities, regulation refs attach.
Drag, split, merge, propagate names to titles in one click.
Outdoor pins on OpenStreetMap, severity-coloured, numbered.
Floorplans, sketches, schematics — same pin language.
Captions drafted, visible text extracted, all on your laptop.
One click. Title, map, legend, area sections. Auto-numbered.
Same drag-and-drop, same keyboard, same firm-branded output — whether it's a 5-photo retainer or a 5,000-photo survey.
No upload. No login. No internet required after install. AI features run locally.
The photos stay where you put them. The app reads them in place.
Tags ride along in the photos themselves (EXIF + XMP). If TAR vanished tomorrow, the work is still there.
TAR's sister app FieldLog already runs on your phone for field forms and signatures. The next phase wires them together — a photo taken on site is auto-uploaded, AI-tagged, GPS-grouped, and queued for the report without ever touching a laptop.
You walk back to the truck. The draft report is already on your screen — codes filled in, areas grouped, photos placed, plan-pinned. You review for 15 minutes instead of building for 4 hours.
TAR doesn't replace anything. The iCloud-to-server flow, the Explorer folders, the Word templates — all still work, untouched. Nothing about how the firm operates today depends on TAR being adopted.
Use it on the inspection where the time savings matter. Skip it on the one where they don't. Try it once. If it doesn't fit how you work, the old workflow is right there waiting.
Same deliverable. Faster turnaround.
Consistent output across every inspector.
Findings traceable to the regulation, every time.
It's already set up and waiting on your laptop. No install, no logins, no cloud accounts — just a folder of photos and a keyboard.
Pick the next inspection on your desk — a five-photo retainer check is the fastest way to feel the difference. We can sit down together and walk through it the first time, then it's yours.
The current workflow stays exactly where it is. If the new one fits better, you'll know inside an inspection or two.
Questions, feedback, or want a walkthrough? Talk to Elliott.