MeshSat Docs

Appearance

Field intelligence ​

A bridge in the field is not just a relay. It watches its own links, its radio environment and the people carrying its nodes, and it acts on what it sees. These are the features that do the watching.

Geofences ​

A geofence is a polygon and a rule about crossing it.

bash
curl -X POST http://bridge:6050/api/geofences \
  -H 'Content-Type: application/json' \
  -d '{
    "id": "base-camp",
    "name": "Base camp perimeter",
    "polygon": [{"lat":60.39,"lon":5.32},{"lat":60.40,"lon":5.32},{"lat":60.40,"lon":5.34}],
    "alert_on": "exit",
    "message": "{node} has left base camp"
  }'
FieldNotes
idrequired, yours to choose
polygonat least three vertices, ordered
alert_onenter, exit, or both, defaulting to both
messagethe alert text

The monitor tracks each node's inside or outside state per zone and fires only on the transition, so a node sitting on the boundary does not generate a stream of alerts. GET /api/geofences lists zones and DELETE /api/geofences/{id} removes one.

Spectrum monitoring ​

With an RTL-SDR attached, the bridge scans the bands its own interfaces use and compares them against a learned baseline. It reports one state per band:

StateMeaning
clearat or near baseline
degradedsustained elevation, not yet characteristic of interference
interferencesomething is transmitting in your band
jammingthe pattern looks deliberate
calibratingstill building a baseline
disabledno scanner

The bands ship pre-tuned to the interfaces they matter to: LoRa EU868 for mesh_0, APRS 144.8 MHz for ax25_0, GPS L1 for gps_0, LTE for cellular_0.

This is not only a display. A band in the jamming state forces its interface's health score to zero, which is how the router learns to stop trying. Endpoints: /api/spectrum/status, /api/spectrum/hardware, /api/spectrum/history, and /api/spectrum/stream for a live feed.

Channel health ​

Every interface carries a composite score from 0 to 100, blending signal, delivery success, latency and cost, with jamming and a deaf receiver forcing it to zero. GET /api/interfaces/health returns them.

The scores are diagnostic: they tell you what your failover priorities should be, and they are where an interface that is quietly failing shows up before anyone notices the missing messages.

Opportunistic satellite burst ​

A satellite pass is short and an SBD frame is 340 bytes whether you fill it or not. The burst queue holds low urgency messages and packs several into a single frame when the window opens, using a simple length prefixed format so the far end can unpack them.

Defaults are ten pending messages and a thirty minute maximum age, whichever comes first. GET /api/burst/status reports pending, size and age limits; POST /api/burst/flush sends immediately without waiting for either.

Pair it with the pass scheduler, which is what tells the queue that a window has opened.

Mesh topology ​

GET /api/topology returns the mesh as a graph: nodes with battery, position and last seen, links with the signal-to-noise ratio measured on them, and totals including the average SNR across the network. A node counts as online if it has been heard in the last 15 minutes. The web UI draws it as a force directed graph and refreshes every 30 seconds.

It answers a question no table answers well, which is whether your mesh is one network or two, and which single node the rest of it depends on.

Dead man's switch ​

Not currently a safety mechanism

The switch arms, counts down, and reports triggered through GET /api/deadman. It does not send anything when it fires, and only saving its own configuration resets the timer, not your activity. The Settings page describes behaviour the Bridge does not yet have.

Do not rely on it. This is tracked as MESHSAT-996 and this page will be updated when it is fixed. For a dead man's switch that does escalate today, use the Hub's SOS and escalation.

POST /api/deadman takes enabled and timeout_min (default 240, minimum 1). GET /api/deadman reports enabled, timeout_min, last_activity and triggered. The switch checks itself once a minute and fires at most once per arming.

What a manual SOS does ​

POST /api/sos/activate is the path that works, and it is deliberately blunt. It sends a broadcast on the mesh and a message through the Iridium gateway, three times at 30 second intervals, going directly to the transports rather than through the router. No access rule, rate limit or transform can intercept it. When the Hub link is down it also sends a Hub uplink frame over satellite or SMS, so the Hub raises the alarm too.

POST /api/sos/cancel stops the repeats. GET /api/sos/status reports whether it is active and how many sends have gone out. Every activation is written to the signed audit log with the trigger that caused it.

Made by MeshSat