Evidence before assumptions
Every finding links back to its source photo and its position on the orthophoto. AI proposes; qualified people decide.
Upload drone photos of a facade. Inspek builds the facade orthophoto, flags and measures defects with AI, and turns the reviewed results into a concise report your client can open without logging in.
Let’s talkOnce photos arrive, Inspek checks photo quality, coverage, and alignment, builds the facade orthophoto for human acceptance, then moves on to AI detection and review. When a gate fails, it opens a recapture task, so the problem never reaches the report.

Inspek creates a Google Drive folder for each facade; paste the link to sync. It checks photo quality, coverage, and alignment. Exclude or restore photos, and turn any gap into a recapture task.
Explore the platformThe first page lists high-risk findings. Each defect carries its photo evidence and measurements, with the full list as a CSV. Clients scan the cover QR code or open the share link to see the orthophoto, photos, and defect locations online. No login, revocable at any time, and shown under your company name.
See what you deliverBased on the current product. Report length varies with project size. AI detection covers a subset of the defect codes; reviewers can apply all of them.
The facade workflow is fully available. Bridges and roofs start with an evaluation on your imagery.

Residential and commercial facades. Review hollowness, spalling, cracks, corrosion, efflorescence, and more under one set of defect codes, from orthophoto to report.

Focus inspection review on cracks and exposed-rebar corrosion, with additional classes in development.

Explore RGB and thermal inspection workflows. Discuss your asset type and validate model fit before deployment.
The facade inspection workflow is available today. Bridge and roof models and developer APIs are still on the way. We always mark which is which.
See what’s available and what’s nextEvery finding links back to its source photo and its position on the orthophoto. AI proposes; qualified people decide.
If photo quality, coverage, or orthophoto acceptance fails, the job stops and asks for a recapture instead of guessing.
Link a job to its building, and defects are matched across inspections and tracked from open to closed.
Inspek is software. We don’t book flights or field inspections. We publish a capture specification so your own pilots or drone service provider can collect photos that work in the workflow.
The full facade workflow is available: photo intake and quality gates, facade orthophoto and acceptance, AI detection with measurement, the review workspace, concise reports, no-login share links, repair marks, and building history. Thermal hollowness detection and the bridge and roof detectors are in beta. The workspace and reports are currently in Traditional Chinese.
No. What matters is the photos. Reference values: forward overlap of at least 80%, side overlap of at least 60%, about 10 mm per pixel (a DJI M4T at 25–27 m), at least 15 RGB photos per facade, and a camera angle within 45° of the wall. To pair thermal with RGB, capture both on the same flight.
As a concise PDF report and a no-login share link. The link opens the overview, the orthophoto, and the photos. It is shown under your company name and can be revoked at any time. The QR code on the report cover opens the link for that report version.
No. Computer vision highlights potential findings for qualified review. Image quality, capture conditions, and the model’s training scope all affect results. Safety and maintenance decisions need professional judgment.
You can export a findings CSV, defect GeoJSON (WGS84), Pix4D camera CSV, and COCO annotations, and import orthophotos produced in Metashape. A public developer API is still planned; tell us about your systems to discuss integration.
Tell us about your buildings and how you capture them. If you can, bring photos from a real job, and we’ll walk through the workflow on your own data.
Let’s talk