Leo · 2026-11
An off-grid security camera with local storage records onto hardware at the site, powered by the site’s own supply, with no cloud account and no upload dependency anywhere in the chain. The 2026 build order is fixed: power budget first, storage second, detection third — and buyers who reverse that order spend twice.

Remote sites — farms, construction yards, cabins, gate houses — break the assumptions of ordinary camera installs. There may be no wired internet, the power supply may be solar or generator, and a stolen camera must not mean stolen evidence. This guide walks the site-planning order that works for off grid security camera local storage, with worked numbers for power and storage.
Cloud camera architectures assume an upload link. Off-grid sites either lack one or share it with everything else on the property, and the failure is silent: cameras keep working while recordings never arrive. A local recording appliance inverts the dependency — recording happens on-site compute, and remote review is a convenience layer rather than the core. The published capability of the woCyber S1 is exactly this pattern: PoE cameras over ONVIF, on-device person, vehicle, and pet detection, 100% of compute performed locally, and recordings on drives you control.
The second argument is evidence integrity. At a remote site the camera and the recorder may share one enclosure; if both are taken, the recordings go too. Planning a second copy — a second appliance in a locked building, or removable media rotated into a safe — is part of the storage design, not an afterthought.

Power over Ethernet simplifies off-grid wiring: one cable per camera carries data and power from a PoE switch or recorder. The budget arithmetic uses the standard’s own numbers, not vendor claims. Power sourcing equipment (PSE) reserves up to 30-W per port for a class-4 device, while a powered device is guaranteed up to 25.5-W; the difference between the two — 4.5-W per port — is the overhead the supply side must absorb. Ten camera ports therefore require planning headroom of roughly 45-W beyond the cameras’ own draw before you count the switch and recorder.
Worked example (illustrative, solar sizing): ten cameras at class-4 planning values plus switch and recorder overhead lands near 200-W of continuous planning load. Over 24 hours that is 4.8 kWh per day. A solar array in a northern-latitude winter may deliver three to four effective sun-hours, so the array needs roughly 1.5 kW of panel to cover the load with battery for the dark hours. The U.S. Department of Energy’s solar program publishes the insolation context behind that math. Site-specific numbers vary; the point is the order — compute the load, then buy panels, never the reverse.

Off-grid storage is finite and expensive to expand, so event filtering earns its place here more than anywhere. Worked example (illustrative): ten cameras at 4 Mbps continuous produce 4 Mbps × 86,400 ÷ 8 ÷ 1,024 = 42.2 GB per camera per day, or 422 GB for the fleet — about 12.7 TB per month. A local appliance with on-device detection that records events instead of continuous footage cuts the figure by the fraction of time detection fires. If events occupy 10% of the day, monthly recording drops to about 1.3 TB, which fits a modest drive set and a small battery budget.
The published S1 capability set includes on-device person, vehicle, and pet detection with a stated false-alarm rate below 3% — the property that makes event-only recording trustworthy at an unattended site. The published S1 retention tiers include 7, 14, and 30 days; retention options available on each model are listed on its product page. The trade-off to plan for: event-only systems record nothing between triggers, so where continuous coverage is required — a gate, an approach road — size those cameras at the full continuous rate and event-record the rest.
| Planning factor | Continuous recording | Event recording with on-device detection |
|---|---|---|
| Daily volume, ten 4 Mbps cameras | About 422 GB | About 42 GB at a 10% event fraction |
| Monthly drive need | About 12.7 TB | About 1.3 TB |
| Power impact | Storage always spinning; larger supply | Same camera power; smaller storage, smaller supply |
| Coverage between events | Full timeline | None — choose deliberately per camera |
| Review burden | Hours of timeline to scrub | Clip list, pre-filtered by detection type |
| False-alarm exposure | None — nothing is filtered | Depends on on-device detection quality |
Plan remote review on the weakest acceptable link. A gate-house site may justify a point-to-point wireless bridge to the farmhouse; a construction yard may rely on a cellular router for clip alerts only. The design test is simple: unplug the internet and verify that recording, detection, and retention continue unchanged. If anything in the system requires the cloud to function, it does not belong in an off-grid design.
Physical-security planning resources from CISA frame the same principle from the risk side: the system’s evidence path must survive the events it is built to detect.
Consider an illustrative orchard: one gate camera on an approach road 400 meters from the buildings, two cameras at the cold-storage loading dock, and no wired internet beyond the farmhouse. The owner plans a mixed design: the gate camera records events only, powered over its own PoE run with detection filtering out wind-blown branches; the loading-dock cameras record continuously because loading disputes need a full timeline. Recordings land on a local appliance in the farmhouse, with a second appliance at the cold-storage building holding a copy of the dock footage. The trade-off accepted: the gate camera’s timeline has gaps, which the owner prices as acceptable against the cost of a larger solar array. That pricing decision — per camera, not per site — is the discipline that keeps off-grid budgets honest.
A working off grid security camera local storage design passes five written checks before any hardware ships:
woCyber publishes the S1 security kit for these builds, with a published MOQ of 100 units per SKU for OEM configuration. The company has shipped 500K+ units over 15+ years, manufactures in Shenzhen with R&D in Beijing, and products are certified CE, UKCA, FCC, RoHS, REACH, and WEEE.

Off-grid systems fail slowly, through drift rather than breakdown, so the plan needs its own calendar. Monthly: verify the battery bank’s charge state and clean panel surfaces; dust on a panel is a quiet capacity cut that shows up as autumn failures. Quarterly: run the restore drill — pull a clip from the archive, open it, confirm the detection labels match the content — and inspect enclosure seals at camera mounts, where a failed gasket turns one winter into one dead camera. Annually: re-run the power budget against actual load, because cameras get added and each addition moves the arithmetic; and re-verify the second copy rotation, since the drive that carried the off-site copy for a year without a check is a hypothesis, not a backup.
None of these tasks is large; all of them are invisible until the night they are needed. A one-page calendar on the site folder — five lines, with initials per completed item — is the difference between an off-grid system that works for years and one that quietly became decoration.
Can cameras run entirely without internet? Yes, if the recorder processes and stores locally. Confirm in writing that no function requires a cloud account, then test with the link unplugged.
How do I power cameras on a site with no mains? A solar array with battery sized from the PoE planning load, or a generator-fed supply. Compute the load first; the panels follow from it.
What stops false alarms filling the drive? On-device detection that classifies person, vehicle, and pet before recording. A stated false-alarm rate below 3% is the published S1 figure; verify behavior on your site.
How long can recordings be kept off grid? As long as the drives and retention policy allow. The published S1 tiers include 7, 14, and 30 days; larger drives extend the window.
What if the recorder is stolen? Keep a second copy of critical footage on another appliance or removable media in a locked building. Design it in from the start.
Build the plan in order — power, storage, detection, then connectivity — and price each camera’s coverage choice individually. The guides below cover the power math and storage sizing in depth.
poe-camera-power-budget-planning
nvr-storage-planning-for-ip-camera-projects
security-camera-system-without-monthly-fees-2026-guide
S1 combines an edge AI NVR and private-cloud NAS in one appliance: person, vehicle and pet detection with a false-alarm rate below 3%, 100% local computation and no subscription.
Browse all platforms →A NAS backup strategy for business rests on three pillars in 2026: a copy that is separate from the primary array, a…
November 20, 2026A security camera system without monthly fees records to storage you own, on hardware that processes video locally, instead of uploading clips…
November 20, 2026PoE cameras win on power stability, bandwidth certainty, and independence from your Wi-Fi; WiFi cameras win on install speed where running cable…
Ask our export team for the OEM/ODM cooperation guide and a written quotation.