Frequently asked questions

Parks and reserves, researchers, and land managers running camera-trap monitoring, and farms, vineyards, and fisheries protecting crops, stock, or ponds from wildlife damage — from a single site to a multi-camera grid.

You do. Footage and detections stay yours, and you can export them anytime. European-hosted by default.

No. It's software you run for your own program, with your own projects and cameras. No membership, no network.

By default, on Untamed Data servers in the Netherlands. You own your footage and detections and can export them at any time. Organizations that need stricter control can run a self-hosted deployment.

You request access rather than signing up openly, which keeps the data clean. Tell us about your monitoring from the Test the platform page, and we send you a login link. You evaluate on your own cameras first; then we prepare a written proposal from your program. See access & costs for how that works.

Most wildlife cameras work. The simplest start is to upload footage yourself — from a computer or an SD card, nothing to set up. Cameras can also upload automatically over FTPS (preferred) or plain FTP; FTPS encrypts your login and files in transit, so use plain FTP only if FTPS is unavailable. See supported cameras for protocols and brands.

Uploads are processed by our European-hosted inference service: animals are detected and species labeled with a confidence score. Uncertain results go to a human review queue. Details on the Wildlife AI page.

Yes. We offer self-hosted deployments and optional white-label / OEM branding for partners and enterprise. See self-hosted & white-label or partner enquiry.

Yes. Beyond software, we offer consultancy for data management, camera selection, placement, and full monitoring programs.

Yes. Use camera grids, species labels, maps, and email alerts for priority or high-value species so protection teams get timely evidence. See the anti-poaching use case, or combine with consultancy for field placement and program design.

Yes. Farms, vineyards, and fisheries use it to see which wildlife is active, get alerted in time, and keep an auditable record for a damage claim or a permit. One honest point: camera footage documents that an animal was present, not that a particular animal caused a particular loss — presence with provenance is what a claim rests on. See crop & vineyard protection or fish-farm predator detection.

Photos are analyzed within seconds; video takes longer — a few minutes for a clip — because many frames are sampled. It all runs automatically in the background. More in photos, video & how the AI reads them.

Photos — especially short bursts — are the workhorse: fast, affordable, and clean per-trigger counts, and more than enough for reliable species ID. Reach for video when behavior or moving groups matter. See photos vs video.

Clear vegetation in front of the lens, face the camera north or south to avoid sun glare, and tune the sensor sensitivity — the AI filters what slips through. See preventing false triggers.

Yes — over any connection you bring, including satellite links like Starlink, which open up sites with no 4G. Truly off-grid needs power for the dish. See Starlink & remote connectivity.

Around 50–80 cm for mixed species — higher for large animals, lower for small ones — angled slightly down. See placing a wildlife camera.

Yes — organize them as projects, each with its own cameras, a shared map, and one dashboard. See multiple cameras in one project.

Yes. Cheap cameras often save money on the optics, and a soft, low-detail image gives the AI less to work with — more uncertain IDs to review, especially at distance or at night. The platform is camera-agnostic, but better optics mean cleaner results. See lens quality and the AI.

Most fire on a passive infrared (PIR) sensor that reacts to moving heat — very low power, but it can miss cold or distant subjects and reacts to heat shimmer. Some use image-based motion detection, which catches more but drains the battery faster. Either way, the cloud AI filters false triggers. See how cameras trigger.

In the European cloud. The camera stays simple — it captures and uploads — and the AI runs on our infrastructure, where the model is more capable, every detection is stamped with the model name and version, and the model can be upgraded without touching your cameras. On-camera (edge) AI trades model power for lower battery and data use; we're exploring that separately for low-power, multi-camera sites (the Smart Wildlife Box), while the platform's full recognition stays in the cloud. See AI on the camera vs the cloud.

Only members of your organization can see your projects, cameras, and detections. See the Privacy Policy for the full details.

Wildlife cameras also record the occasional walker, ranger, or passing car. By default the platform detects people and vehicles and blurs them automatically before the footage is stored — the whole body, not just the face — and the blur can't be undone. The one exception: the untouched original is kept for seven days in case an animal is blurred by mistake, then permanently deleted. For video, the whole clip is blurred and the audio track is removed. It's a safeguard, not a guarantee — someone small, distant, or half-hidden can be missed, and if a person fills too much of the frame the capture is flagged for you to decide rather than blurred. An organization administrator can turn it on or off. More in blurring people & vehicles.