RouteViewNet is a local-first network and memory health monitor for Linux.
It watches your connection and your computer around the clock, scores their
health from 0 to 100, and explains problems in plain English.
No cloud. No account. No telemetry.
Free and open source (MIT) · Installs in one command · Runs entirely on your machine
0health
localhost:4545
Overview
Health score ?96✓ healthylast 24h
Download · wlp2s0 ?1.8 MB/stotal 458 MB
Upload · wlp2s0 ?890 KB/stotal 61.0 MB
Devices seen ?3visible from this machine
Open alerts ?0all clear
Gateway · 192.168.0.1 ?✓ ok2.4 ms
Internet ?✓ ok18.2 ms
DNS ?✓ ok6 ms
Memory ?34%10.7 GB of 31.2 GB
Load average (1m) ?0.868 cores
Daemon memory ?33.5 MB22 goroutines
▼
"the Wi-Fi is slow""pages will not load""is it the router?"
"The Wi-Fi is slow" has a dozen possible causes: your router, your provider,
the DNS address book, a bad cable, or a computer that is simply out of memory.
Guessing means restarting things at random. RouteViewNet measures each link in
the chain continuously, so when something breaks you can see exactly which step
failed and what to do about it.
What it does
One score that means something
A deterministic 0 to 100 health score computed from every check: router
reachability, internet reachability, DNS speed, packet loss, memory, swap,
and processor load. 90 or above is healthy; every lost point is explained.
The three lights: Gateway, Internet, DNS
The three-step path every web page takes, each checked every few seconds
with response times. Which one fails tells you where the problem lives: your
home network, your provider, or the address book.
Live bandwidth, per connection
Download and upload rates for every network interface, with automatic
detection of the primary one on multi-homed machines. History from 15
minutes to 7 days, plus error and drop counters that expose bad cables,
weak Wi-Fi, and overloaded links.
Data usage by app
See which programs moved the most data in any time window, with a view of
every app that used 20 MB or more. Counted from the kernel's own connection
counters and attributed to the owning process.
Internet health checks
Continuous latency and packet-loss probes to well-known reference
addresses (configurable), plus DNS lookup timing for both IPv4 and IPv6
records, charted over time.
Devices on your network
Every device your computer talks to, with vendor lookup, optional
hostnames, nicknames you assign, and a trusted flag. An unfamiliar device on
your Wi-Fi raises a gentle alert.
Memory and load, demystified
System RAM (counted honestly: Linux cache does not count as used), swap
pressure, 1/5/15-minute load against your core count, and the top memory
consumers by name, so the culprit is one glance away.
Alerts that do not cry wolf
A problem must be seen twice in a row to open an alert, and alerts
resolve themselves when the signal recovers. Debounced, deduplicated, and
never requiring a dismiss button.
A Troubleshoot page that talks like a person
Rule-based diagnosis of the currently open alerts: a one-line likely
cause, the evidence behind it, and ordered suggested actions. "Your gateway
is fine, DNS is slow; check your resolver."
Tune it without restarting
Every knob lives in the dashboard: alert sensitivity (memory percentage,
DNS latency, packet loss), check intervals and targets, per-collector
toggles, data retention, and privacy switches. Changes apply the moment you
save; the few that need a restart say so explicitly.
A real desktop app
RouteViewNet lives in your applications menu with its own window and
icon, not just a browser tab. If the background monitor is ever stopped,
the app offers to start it with your system's normal password prompt. The
browser dashboard at localhost:4545 keeps working too.
The monitor watches itself
The daemon reports its own memory footprint in the dashboard, so you can
verify the monitoring tool is not the thing slowing you down.
01
12
0telemetry calls
0SQLite file
0health score
Where exactly did it break?
Gateway2.1 ms
Internet18.4 ms
DNS11.7 ms
All three steps healthy. Watching.
Who used the data?
Per-app usage in any time window, straight from the
kernel's connection counters.
firefox
412 MB
steam
268 MB
spotify
141 MB
apt
96 MB
code
63 MB
slack
41 MB
View all with 20 MB or more →
Up and running in a minute
1
Install.
One .deb package. The background monitor starts immediately and
survives reboots.
2
Open.
Launch RouteViewNet from your applications menu (a native desktop
window) or visit localhost:4545 in a browser.
3
Understand.
Glance at the score. If something is wrong, the Troubleshoot page
tells you where and what to try.
$
Your data never leaves your machine.
[ ] No cloud, no account, no telemetry of any kind.
[ ] Everything lives in a single SQLite file on your disk; history older than 7 days is cleaned up automatically (configurable).
[ ] Process command lines are not stored by default, since they can contain secrets.
[ ] The dashboard answers only on the computer itself. Opening it to your LAN is an explicit setting with a persistent warning.
[ ] Defense in depth: host-header validation, WebSocket origin checks, strict CORS, a sandboxed systemd service with least-privilege capabilities.
nothing exits the loop
Built for Linux users, honest by design
THE STACK
A single Go daemon, SQLite storage, a React dashboard embedded in the
binary, systemd-managed. Hot-reloadable settings from the dashboard.
HONEST LIMITS
Shows devices this machine talks to, not a full LAN scan. Per-app usage
counts direct (TCP) connections only. Load average only for CPU. No
authentication in v1, so keep it local.
Questions, answered plainly
Does it slow my computer down?▾
No. It typically uses less memory than one browser
tab, and it proves that on its own System page.
Do I need to keep the window open?▾
No, the monitor runs in the background; the window
is just a viewer.
What Linux versions?▾
Debian and Ubuntu with systemd; Go 1.24 or newer to
build from source.
Is it really free?▾
Yes, MIT licensed.
Can I see it from my phone?▾
Not out of the box; v1 has no authentication, so it
stays local by default.