Mesh-Client supports emergency communications (EMCOMM) on Meshtastic, MeshCore, and Reticulum. This guide is for operators: how emergency reports are sent and received, what the Incident tab shows, and which settings matter. For problems, see Troubleshooting — MECP.
MECP (Mesh Emergency Communication Protocol) is structured emergency text sent as ordinary chat:
MECP/<severity>/<codes> [freetext]
| Severity | Meaning | Alert |
|---|---|---|
0 |
MAYDAY | Siren tone and emergency toast; ignores mute, even in a focused chat |
1 |
URGENT | EAS-style tone and toast; ignores mute |
2 |
SAFETY | Short tone when not muted |
3 |
ROUTINE | Short tone when not muted |
- Codes are a letter plus two digits (for example
M01). Labels come from the MECP language packs, so codes display in your language. - A report is at most 200 bytes (UTF-8).
- Drill codes (
D01,D02) mark a report as a drill: it is listed but never plays an alert tone or toast and never adds an unread alert badge. Open MAYDAY and URGENT drills still count toward the Incident tab badge. - Tones and volume are configurable per severity in App → Notifications → Notification tones. MAYDAY and URGENT keep their mute bypass and a 10% volume floor. See Notification sounds.
- Chat channel and DM chips show a shield icon, colored by the most severe unread MECP report in that conversation.
- Turn on App → MECP → Show MECP buttons in Chat (off by default).
- In Chat, open the conversation to send to, then click the red siren button to request help or report conditions, or the blue shield button to respond.
- Choose severity and codes, add free text and optionally your position, and send. Both workflows start at ROUTINE with no codes selected. The response workflow opens on acknowledgements and help updates. Use Show all MECP codes if you need a code from another category.
Both workflows use the same standard MECP codes and translated labels, so other clients can read them. On MeshCore, queued reports and retries use the destination channel’s saved scope. Choose Unscoped beside Send for mesh-wide floods, or a saved region for local floods; Default follows the companion radio’s default.
Every report goes through the emergency outbox. It is sent right away when the radio is up, and it keeps retrying (after 30 seconds, 1 minute, 2 minutes, then every 5 minutes) until the network acknowledges it. Acknowledgement means the radio heard the mesh repeat or confirm it (a Meshtastic ACK, a MeshCore repeater rebroadcast or DM ACK, a Reticulum delivery receipt), or another station acknowledged or relayed your report. There is no age cutoff or attempt limit. While a report is waiting, Chat shows it with Waiting for network acknowledgement… or a retry countdown. If a report is stuck, use Stop retrying on it (you are asked to confirm). A soft cap of 20 queued emergency reports blocks the least urgent extras, but never deletes anything and never blocks a MAYDAY.
The Incident tab is always visible on all three protocols, at the bottom of the rail next to App. Its red badge counts open MAYDAY and URGENT reports (drills included), so you notice them even if you rarely open the tab.
Each row shows severity, sender, MECP codes, free text, ACK count, which protocols heard the report, and whether a distress beacon is active. The same report heard on several protocols (for example bridged from Meshtastic to MeshCore) appears once.
Coordinates come from the report itself or the sender's last known position, and appear on the Map under Layers → Emergency incidents (Meshtastic, MeshCore, and Reticulum maps).
| Button | What it sends |
|---|---|
| Acknowledge | An R01 ACK echoing the report's codes |
| Confirm | A B02 beacon ACK (shown instead of Acknowledge while someone's beacon is active), asking them to reduce beacon rate |
| Resolve | Nothing; closes the row locally |
| Cancel beacon | A B03 "I am OK" on the beacon's protocol and channel, then resolves the row. Shown only for beacons you sent |
ACKs are sent on the protocol where the report was heard (preferring your active tab). They are heard-by-network, best effort: an ACK count means ACK copies were overheard on the mesh, not that the person in distress read anything. Treat it as "the network has it", never as a read receipt.
You can't cancel someone else's beacon: B03 means "I am OK", so only the originator can send it.
The MAYDAY siren and URGENT tone play once, and the toast closes after 10 seconds. Everyone gets that one-time alert. Stations that must not miss a report (an EOC, a net control, a monitored shelter) can turn on App → MECP → Incident Command station (standing alert). It is off by default.
With it on, a new MAYDAY or URGENT report (not a drill) also opens a red banner at the top of every tab, and the window flashes in the taskbar (or the dock icon bounces on macOS) when the app is not focused. The banner stays until you open the Incident tab, press Mark seen, acknowledge, or resolve the incident. View opens the Incident tab.
"Seen" is local to this computer and sends nothing over the mesh. Acknowledging is still a separate, deliberate step. Reports loaded from history at startup, your own beacons, and drills never open the banner. A report that escalates to MAYDAY or URGENT, or reopens after being resolved, opens it again.
For unattended stations, App → MECP → Repeat unseen MAYDAY/URGENT alert (minutes) (under the station toggle) replays the tone at that interval until the alert is seen. It is blank (off) by default.
Anyone on the mesh can send a MECP report, so mesh-client limits how much noise one sender can make:
- Repeat silencing: after 3 alerts from the same sender within 10 minutes, further reports from that sender stop sounding. They still appear in the Incident tab (without the banner), and a one-time notice offers Block sender. If many different stations send reports within a minute, further tones pause briefly as well.
- Block alerts: each row in the Incident tab has Block alerts. Choose Block only, or Block and resolve to also close that sender's open incidents locally. Reports from a blocked sender no longer alert, open incidents, or get bridged by RF rebroadcast. They are still written to the audit log (marked blocked) and still show as text in Chat. Unblock in App → MECP → Blocked emergency senders.
- Block a node completely (Meshtastic and MeshCore): Block in a node's detail view drops all of that node's messages before they reach Chat, so it cannot send MECP either. Unblock in App → Blocked nodes. Reticulum has its own blocklist in the Reticulum network settings.
Node IDs can be spoofed, especially on Meshtastic, so blocking stops casual abuse but not a determined attacker. Only block a sender you are sure is abusing MECP.
Every inbound MECP report is appended to mecp-received.log in the app data folder (separate from the session app log, rotated to mecp-received.log.1). Export it from App → MECP → Export MECP log. The logs are included only in the private developer support bundle, never in the public-safe GitHub report.
App → MECP can bridge MECP reports between a Meshtastic channel and a MeshCore channel. It is off by default.
- Each rule has endpoint A and endpoint B (channel indices 0–7). Reports go A → B by default; turn on Bidirectional to also send B → A.
- Only reports heard over RF are bridged (not MQTT-only), and never your own reports, history, or drills. A loop guard prevents ping-ponging the same report.
- After each bridged report, a short plain-text notice follows:
MECP from <sender> via <Meshtastic|MeshCore> (<channel name>).
When the TAK server is running and a channel is served to ATAK (or direct messages are on for the radio), an MECP report heard there reaches ATAK in words, in the app's language, with its code line kept below it, the same way MeshCore TAK Bridge shows one:
🆘 MAYDAY: Injury · Fracture / immobile · 2 people · 48.6520, 20.1305
MECP/0/M01 M07 2pax 48.6520,20.1305
A drill starts with DRILL. Anything else is passed on to ATAK as it is.
Each open incident with a position, the same ones the Map shows under Layers → Emergency incidents, is also a spot marker on every connected ATAK device's map: red for MAYDAY, orange for URGENT, yellow for SAFETY, white for ROUTINE, grey for a drill. It is named with the severity and the sender (🆘 MAYDAY Alice), and its remarks hold the report in words, the code line, when it was last heard, where the position came from when it is the sender's last known one rather than the report's, a distress beacon, and the acknowledgements heard. A marker updates as the incident does and is removed from ATAK when the incident is resolved in mesh-client. Cancelling your own beacon resolves it; a cancel heard from another station clears the beacon line, and the marker stays until someone resolves the incident. At most 50 are shown, most severe first. Markers are not sent to a remote TAK server.
Sound ATAK's emergency alarm for an MECP MAYDAY (in the TAK server's settings) also sets off ATAK's own emergency alarm, the one ATAK raises for a 911 alert, on every connected device, the same as MeshCore TAK Bridge's option. Anyone on a channel can send a MAYDAY, so it is off unless you turn it on, and it takes effect at once, whether the server is running or not. It sounds once for a MAYDAY with a position that mesh-client is alerting about too: never for a drill, an URGENT or lesser report, your own beacon, a blocked sender, or a sender whose alerts have been silenced for sending too many, and not for a report read from history or one you had already seen when the alarm was turned on. The alarm is linked to the incident's marker and named for the sender (Alice-Alert). It is cancelled on ATAK when the incident is resolved, its beacon is cancelled, the sender is blocked, or the option is turned off. It sounds on the devices connected when it starts; one that connects later sees the marker but not the alarm. An alarm is sent once and not refreshed: if no cancel reaches it, for example because the server stopped, ATAK drops it 30 minutes after it went off. Alarms are not sent to a remote TAK server.
App → Notifications has ops alerts for nodes you mark as watched:
- Silence: a watched node has not been heard for the configured minutes, with an escalation at twice that time. Nodes never heard are not "silent".
- Battery low: a watched node's reported battery drops below the threshold (default 10%). Fires once, then re-arms after it recovers 5 points.
- Link down: your Meshtastic or MeshCore radio link drops unexpectedly. Never fires on a manual disconnect or while reconnecting. Reticulum is not included (it runs as a separate stack, not an RF link).
- Standing alert and repeat for unseen MAYDAY/URGENT: in App → MECP; see Unseen emergency alert.
- Nodes: NodeList → Export JSON (topology) or Export CSV.
- Diagnostics: Diagnostics → Export JSON (visible rows for the active protocol).
- MECP audit log: App → MECP → Export MECP log.
- Incident log: Incident tab → Export log → JSON or CSV. Includes every stored incident (open, acknowledged, and resolved) with severity, codes, sender, relays, timestamps, coordinates, ACK peers, and beacon state, oldest first. Each incident also carries an
eventstimeline: when it was received, relay copies heard, each ACK heard (and from whom), your own ACK sent, beacon confirmed or cancelled, resolved, and reopened. In CSV the timeline is one cell of semicolon-separated entries. Use it for the after-action record.
- MGRS grid: Map → Layers → MGRS grid (off by default). Grid precision adapts to the zoom level; labels appear when few enough squares are visible.
- Measure: the Map measure control draws a polyline and shows its length.
- USGS Topo basemap (US only) is available and can be cached for offline use like the other basemaps. See Troubleshooting — Map tab without internet.
- Position history for senders of open or acknowledged incidents is kept even when normal history pruning runs, so their tracks stay available until the incident is resolved.