A fleet console can tell you a machine stopped reporting. It can't tell you the camera is on a port that still has power, that the uplink to the far building is degraded, or that there's a device on the network nobody ever installed. NetYard is the physical layer — the authoritative record of what's actually in the building — one dashboard per site, and the site feed behind NetFoundry.
The physical layer. What's actually plugged in, powered and reachable at this site.
Agents report on machines. Nothing reports on the place — the cabling, the switches, the radios, the uplink, and the forty devices that have no agent and never will.
Internet → modem → router → switches → access points → bridges → servers, in the order you actually troubleshoot, each with live state and a green / amber / red summary.
Every device at the site, named, typed, grouped and searchable — including the ones with no agent, no owner and no label.
Any device drawn as the chain of equipment it depends on, from the internet down to its port, with live status on every link.
Switches as port diagrams: what's connected, which ports are powering devices, how much budget is left, and what moved since last month.
The uplink as a first-class device — provider, connection type, speedtest history, and a dated record of every address the ISP has handed you.
Optional. Access schedules, screen time and media activity for residential sites — switched off entirely at commercial ones.
Every device sits at the end of a chain of physical equipment — the uplink, the modem, the router, the switch, the port, the radio. NetYard resolves live status for each link from the best source it has. The degraded link is the answer, before anyone gets in a car.
Every link in that chain is read from equipment already installed at the site. Nothing is installed on the device being diagnosed — most of them couldn't run an agent if you asked.
You don't draw the topology. NetYard works out which switch and port each device is on, and which switch links to which, from what the equipment reports about itself.
When a link can't be read because credentials are missing or a device isn't answering its management interface, it says exactly that — it never reports "not found" for a permission problem.
Each device shows which of its management interfaces are currently responding, so an empty panel is explained on the card instead of becoming an investigation.
"Powered but silent" and "no power at all" are different faults with different fixes. A tool that only pings can't tell them apart; the port tells you which one you have.
Two layers of networking, a switch that misreports its own port numbering, a firewall that names interfaces nothing like their physical ports, radios in bridge mode — all handled, because each one broke it once and was fixed.
A sleeping laptop won't answer a scan. A camera on another part of the network won't appear in a local sweep. A modern phone changes its hardware address weekly and looks like a new stranger each time. NetYard asks everything that already knows the network and reconciles the answers, so the list in front of you is actually complete.
Which device is on which port, which ports deliver power, and how the switches connect to each other. This is the source a scanner can never replace.
Devices on other parts of the network that a local sweep cannot reach, plus what each firewall interface is actually configured for — address range, description and all.
Assignments and reservations from AdGuard Home, EdgeRouter, Cisco ASA or MikroTik — including the names people typed when they joined, which beat anything a scanner can guess.
Who is connected to which access point, on which network name, with signal and link quality — and for wireless bridges, the wired devices living behind them.
New devices appear the second they join rather than at the next sweep, along with roaming, disconnects and unstable-port events as they happen.
NetYard works out whether the site sits behind a second layer of networking and, if so, finds the ISP modem and router above you — the equipment that quietly breaks remote access and that most tools never see.
Apple TVs, Sonos speakers, Roku and Chromecast players, Samsung and LG televisions, Hue hubs, smart plugs, media servers and printers are recognized by make, model and the name their owner gave them.
Confident identifications fill blank fields; anything less certain waits behind a badge for approval. Reject one and the same source never offers it again. Nothing you typed is ever overwritten.
A laptop's wired and wireless connections, or a virtual machine and its host, collapse into a single entry instead of three unexplained strangers.
Location, brand, model, asset code, notes, a photo, a QR code, color-coded groups, and whether this one is worth an alert. Search runs across all of it.
The type isn't decoration. Each brings its own icon, card layout and defaults — access points don't show a disk gauge, the modem opens a signal panel, and the internet itself is a device you can watch.
The scan knows what answered. The switch knows what's plugged in. The address server knows what asked to join. The Wi-Fi knows who's associated. None of them knows what the others are missing. NetYard correlates all of them into one row per device — and the interesting rows are the ones with a gap in them.
| Device | Answered scan | On a switch port | Has an address lease | On Wi-Fi | Verdict |
|---|---|---|---|---|---|
| office-nas01 | ✓ Yes | ✓ Port 14 | ✓ Reserved | — wired | ✓ Healthy |
| jdoe-macbook | ✕ Silent | — not wired | ✓ Active | ✓ Associated | ◦ Asleep, not down |
| unknown · 10.10.1.84 | ✓ Yes | ✓ Port 6 | — no lease | — not on Wi-Fi | ✕ Unrecognized |
| gate-camera-03 | ✕ Silent | ! Port 9, powered | ✓ Reserved | — wired | ✕ Cable or device |
| ap-patio | ✓ Yes | ✓ Port 11 | ✓ Reserved | ✓ 8 clients | ✓ Healthy |
A sleeping laptop is not an outage. NetYard checks in a way that tolerates power-saving network cards and slow-waking machines, because an earlier build reported those as down and people stopped reading the alerts — which is worse than no alerts at all.
A guest's phone leaving isn't an incident. Mark a device as not-alerting and it reports gray for information rather than red for alarm — visible, but not in the number you're judged on.
A device on a switch port with no lease and no name is flagged as unrecognized rather than quietly listed. Devices you've deliberately decided not to care about can be hidden for good, so the list stays worth reading.
Several manufacturer-based guesses were removed for being too ambiguous — one maker's chips ship in thousands of unrelated products, and a confident wrong label is worse than an honest blank. NetYard reads from your equipment; it doesn't reconfigure it.
That question ends most site calls, and almost nothing answers it. NetYard draws each switch as the port diagram you'd sketch on a whiteboard, and each port names what's connected to it.
A device visible both on its own port and on a shared link between switches is shown on the port it's genuinely plugged into — not wherever it happened to appear last.
What each port is drawing and what the switch has left — so you know before you commit to the ninth camera on an eight-camera budget.
One button checks everything connected to a port in turn, separating a bad cable run from a bad device without unplugging anything.
When a device turns up on a different port next month, that change is logged — which is how you find out somebody re-patched the rack and didn't say.
Per-access-point client lists with signal, noise and link quality, plus point-to-point link health between buildings — the thing that degrades for weeks before it fails.
A port or a wireless client that keeps dropping and reconnecting never trips an outage alert. NetYard notices the pattern and says so — and ignores the normal roaming that looks identical if you're not careful.
NetYard treats the connection as a device in its own right, pinned to the top of the dashboard and graded continuously against several independent providers, so one company having a bad afternoon can't masquerade as your outage. When the client says "the internet is slow", you answer with data instead of their impression of it.
Download and upload measured automatically every half hour and on demand, charted over time. "It's been slow lately" becomes a graph with dates on it.
Gateway, upstream and name-resolution checks as an indicator grid, measured against several well-known providers separately — so one provider's problem is visibly not an outage.
A plain-language verdict on whether the site has a normal connection or a shared carrier one. That's the difference between "port forwarding is broken" and "port forwarding will never work on this plan" — a conversation that otherwise takes three support calls.
A dated log of every public address the ISP has handed this site. Invaluable the day a remote connection stops working for no apparent reason.
On cable connections: per-channel signal strength, noise margin and error counts, kept as history, with each warning explained in words. Signal problems become visible before they become outages.
A live feed of which services the network is requesting, how much is being blocked, and whether any device is quietly bypassing the site's filtering. It watches; it doesn't interfere.
NetYard reads from equipment you already own, with read-only credentials scoped per device. Nothing is replaced, nothing is reconfigured, and every integration is optional — add one and that part of the picture fills itself in.
Credentials are stored per device and only ever used to read. NetYard never needs write access to a switch, a router or a firewall to do its job.
Each device's admin page opens from inside NetYard, on NetYard's own address — so a browser never needs to reach the site's internal network, and you stop keeping twenty bookmarks for twenty internal addresses.
Built-in web interfaces on network gear are famously badly behaved when opened through anything else. NetYard ships ready-made handling for Ubiquiti, Cisco, Netgear, MikroTik, Proxmox, printers and phone systems, so they work rather than half-load.
Each integration form explains what it needs and where to find it, and has a test button that reports the real reason it failed. Least-privilege throughout.
NetYard runs at homes as well as offices, and household features work from what the network can already see — so there's nothing to install on a child's tablet and no subscription. At commercial sites a single switch removes all of it, dashboard and all.
Bedtime and homework windows and a daily limit, per child, with their devices assigned. Shared devices follow the house rule and the strictest one wins — so the family tablet can't be unlocked by the youngest sibling's schedule.
Stackable "+30 minutes" and "+1 hour" with a live countdown, and the next scheduled change shown up front so you always know what you're overriding.
What's playing right now on each Apple TV with cover art, plus recently-watched grouped by day with the show name and an age-appropriateness flag pulled in automatically.
This data is governance-sensitive, so it stays at the site. It is excluded from the fleet feed on purpose — household behavior has no business in a multi-client console, and the exclusion isn't a setting someone can flip by accident.
NetYard installs on a machine that stays on at the site — a small server, a spare mini PC, even a Raspberry Pi — and starts finding devices as soon as you point it at the network. Nothing is installed on the devices it monitors.
Everything NetYard collects stays on the machine you installed it on. There is no NetYard cloud to send it to, no vendor account, and no per-device pricing as the site grows.
A low-write mode keeps long-term history off the local card and in storage you choose, holding recent measurements in memory through an outage. Years of life out of hardware that would otherwise wear out in months.
Local history, local alerts, local dashboard. A site keeps monitoring itself when the link between sites is down — no site goes blind because head office is unreachable.
Reporting upward is off until you turn it on, needs a key you generate and can replace, and is read-only by design — nothing about the site can be changed through it.
The site name sits across the header so fourteen open tabs don't become a guessing game, and each site keeps its own timezone, so schedules fire at the right local hour.
Which networks to sweep, how often each check runs, which equipment it talks to and which features are visible are all per-site settings, adjusted in the interface. Light and dark themes throughout.
The meaningful improvements came from running against real networks and finding the output wrong: devices reported down that were merely asleep; a switch family that misnumbers its own ports; a firewall reporting a processor at 100% that wasn't; devices attributed to the uplink instead of the port they were really on. Each traced to a cause and fixed. That's why the port map can be trusted.
Where a device exposes nothing to read, NetYard says so and links you to its own interface rather than inventing a number. Where a guess was too ambiguous to be useful, it was removed rather than shipped — a confident wrong answer costs more than a blank one.
A fleet console tells you which site needs attention. NetYard gives you the one thing it can't: what's actually happening in the building, down to the port.