Skip to content

Releases

MeshHome 1.3.3

Choose a tag to compare

@roach0816 roach0816 released this 18 Nov 21:51

A fix for saving node settings on MeshCore radios.

Fixes

  • Settings → Radio → Configure node settings → Contacts & routing: saving showed "Request failed" when Default flood scope was empty, on firmware that has flood scopes. Nothing was saved. It saves now, including Make room when the contact list is full from 1.3.2. Restoring a radio backup with no flood scope ran into the same problem and is fixed too.
  • If the MeshCore library fails while MeshHome is changing a radio setting, the page now says which setting and points to Settings → Health & logs, instead of "Request failed".

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.3.2

Choose a tag to compare

@roach0816 roach0816 released this 18 Nov 21:03

Choose what your MeshCore radio does when its contact list is full.

Settings → Radio → Contacts & routing → Make room when the contact list is full

  • On: a new contact replaces the one the radio heard from longest ago.
  • Off: new contacts aren't added until you remove one. Nothing is replaced.

Either way, favorites are never replaced, and MeshHome keeps a replaced contact's messages (the contact shows as Removed from radio). The setting also shows how many contacts your radio holds, for example 350.

Good to know: a MeshCore radio can only read direct messages from contacts on its list.

Details

  • For MeshCore radios whose firmware has the setting. It's hidden on older firmware and on Meshtastic radios, which always make room for new nodes by themselves.
  • Saving changes only this setting. The radio's per-type auto-add choices and its max-hops limit stay as they are.
  • Radio backups include it. Restoring an older backup leaves it unchanged.

API: behavior in GET /api/radio/config has overwrite_oldest (null on firmware without it) and the read-only max_contacts. PUT /api/radio/config/behavior accepts overwrite_oldest; leaving it out keeps the current value.

Not yet tested on a real MeshCore radio.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.3.1

Choose a tag to compare

@roach0816 roach0816 released this 13 Nov 15:04

Tracker telemetry, easy to find.

On the Tracking page

  • On each tracker's card: its battery, temperature, humidity and signal, at a glance.

  • Telemetry opens everything the tracker reports. The button is in the tracker list, on the tracker's card along the bottom of the map (one tap on a phone), and in its pin's popup. It shows:

    • Battery: charge and voltage, or "Plugged in".
    • Sensors: temperature, humidity, pressure and light.
    • Signal: SNR, RSSI and the route its last position took.
    • Device (Meshtastic): uptime, how busy the channel is, and its own airtime.
    • History: battery and temperature charts over 6 hours, 24 hours or 3 days.

    Each group says how long ago it was reported, and Something missing? says what to turn on in the tracker's app.

Fixes

  • A Meshtastic tracker's battery now shows as soon as MeshHome connects, from the radio's node list. Before, it waited for the tracker's next telemetry broadcast, every 30 minutes by default.
  • Readings that were dropped are now kept: Meshtastic trackers' humidity, light, uptime, channel use and airtime, and MeshCore trackers' humidity and light.

API: each tracker in GET /api/tracking has telemetry (its latest readings with times) and external_power, and its sensors include humidity and lux. GET /api/tracking/devices/{id}/telemetry?hours=24 returns the history behind the charts.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.3.0

Choose a tag to compare

@roach0816 roach0816 released this 13 Nov 14:26

Settings → Health & logs: what's working and what isn't, and MeshHome's own log.

Health

  • What's checked, every 30 seconds:
    • the radio connection, the radio HAT, and when the radio last heard the mesh
    • sending and collecting messages
    • trackers
    • the database, MeshHome's background services, notifications, the HTTPS certificate and updates
    • on Linux: the computer's disk, memory, temperature, power supply and clock
  • When something isn't working, its line opens by itself with what it means in plain words, numbered steps to fix it, the facts behind it, and links to the setting and to its log lines. For example: "Nothing is accepting connections at 192.168.1.40:4403" rather than "Errno 111", or "The radio HAT's software (meshtasticd) isn't running".
  • Where you'll notice it: a dot beside Health & logs in Settings when something needs attention, and a What's wrong? link in Settings → Radio while the radio is offline.
  • Copy report (Download report without HTTPS) copies the whole check, recent problems and the radio's connection history as text, to paste when asking for help. It contains no passwords or keys, but does include network addresses and names.

Log

  • History keeps warnings, errors and events for 30 days. Events are the radio connecting and going offline, MeshHome starting and stopping, and radio settings, radio HAT changes and updates being requested. A line that repeats, such as the radio retrying all night, is one entry with its count and time span.
  • Since start shows every line since MeshHome started.
  • Filters: part of MeshHome, problems only, and text search.
  • Database outages: lines are kept in memory while the database is unavailable and saved when it's back, so an outage shows up in the history too.
  • Stopped services: a background service that stops after an error is now reported.

API: GET /api/system/health returns the same checks and can be read with an API key, so monitoring tools can watch it. GET /api/system/logs and GET /api/system/health/report are for signed-in browsers. See docs/API.md.

Not yet tested on a Raspberry Pi: the radio HAT and "This computer" checks. They read the same system information as sudo meshhome status.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over; the log history starts with this version.

MeshHome 1.2.2

Choose a tag to compare

@roach0816 roach0816 released this 13 Nov 13:50

A fix for the radio staying offline.

The radio reconnects after a database outage. If PostgreSQL restarted while the radio was reconnecting (as an automatic overnight security update can do), MeshHome stopped trying to connect to the radio, which then stayed offline until MeshHome was restarted. It now keeps retrying, with Settings → Radio showing "Database unavailable" until the database is back, and reconnects by itself. Restarting the radio HAT didn't help before; sudo meshhome restart did.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.2.1

Choose a tag to compare

@roach0816 roach0816 released this 13 Nov 01:15

One place to choose MeshCore or Meshtastic.

Settings → Radio → Radio connection

  • Mesh (MeshCore or Meshtastic) is now the only place to choose. The radio HAT box no longer has its own Switch to … button, and setting up the HAT installs the mesh chosen above.
  • With the radio HAT running the other mesh, Save becomes Switch the HAT to Meshtastic… (or MeshCore). It asks first, with the RAK module and IO slot for Meshtastic, then switches the HAT's software. MeshHome switches with it, as before.

API: POST /api/system/radio-hat takes "use": true to also make the HAT MeshHome's radio, running that mesh.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.2.0

Choose a tag to compare

@roach0816 roach0816 released this 12 Nov 23:03

MeshHome can now run on Meshtastic as well as MeshCore. Choose the mesh in Settings, and contacts, channels, messages, tracking and the rest of the app follow it.

MeshCore or Meshtastic

  • Where to choose: Settings → Radio → Radio connection → Mesh: MeshCore or Meshtastic.
  • What follows it: everything shows that mesh's contacts, channels, messages, notifications, bot replies, map and trackers.
  • Switching back: the other mesh's are kept, and come back when you switch.
  • With the radio HAT, MeshHome runs what the HAT runs: switching the HAT between ZephCore and meshtasticd switches MeshHome with it. The HAT stays selected as the radio either way.
  • Meshtastic radios:
    • The radio HAT running meshtasticd.
    • A Meshtastic node on the network (its API on TCP port 4403, e.g. an ESP32 node with Wi-Fi).
    • A simulated mesh: two trackers on your course, a router at the car, and a hiker who chats and answers messages.

Messaging on Meshtastic

  • Contacts are the nodes the radio has heard, with their node id, role, battery, signal and hop count. Favouriting and removing a node go to the radio.
  • Channels are the radio's own. An unnamed primary channel goes by its preset ("LongFast"), as in Meshtastic's apps. The + in the conversation list adds a channel, or joins another group's from their share link or QR code.
  • Direct messages:
    • The radio encrypts them end to end, and the recipient's acknowledgement shows as Delivered.
    • The radio retries on its own, so each message is sent once.
    • If the recipient refuses it because the two radios haven't swapped keys yet, MeshHome sends this radio's details and the message says to send it again.
  • Length: messages are up to 200 bytes.

Node settings for Meshtastic (Settings → Radio → Configure node settings)

  • The radio itself: name and short name, LoRa region, modem preset, hops and transmit power, role, and a fixed position (also the base on MeshHome's maps).
  • Channels:
    • Name, key (private, or Meshtastic's default), and how precisely positions are sent.
    • Each channel's QR code, to put it on trackers and other nodes. On an iPhone, scan it with the Camera app; on Android, use the scan button in the Meshtastic app's Settings → Channels.

Tracking on Meshtastic

  • Adding a tracker: Meshtastic trackers are added from the nodes the radio has heard, or from a contact, and every position they broadcast is recorded.
  • The channel: for exact positions only your team can follow, give the primary channel a private key with exact positions in Node settings, and scan its QR code into each tracker.
  • From 1.1: the separate "Meshtastic trackers" connection in Settings → Tracking is gone: Meshtastic is now the radio itself. A 1.1 setup carries over: the radio HAT running Meshtastic (or the Meshtastic node you'd set) becomes MeshHome's radio.

Each mesh's own features

  • Hidden on Meshtastic:
    • Repeater and room server administration.
    • Routes, traces and the preferred repeater.
    • Signed messages and contact QR codes.
    • Radio firmware updates and radio backups.
  • Hidden on MeshCore: the Meshtastic settings.

API change: these 1.1 endpoints are removed:

  • GET/PUT /api/settings/meshtastic
  • GET /api/tracking/meshtastic
  • POST /api/tracking/meshtastic/channel
  • GET /api/tracking/meshtastic/channel-url

The Meshtastic radio's settings are under /api/node/meshtastic. GET /api/status reports the technology in use as radio_config.tech, and MeshCore-only endpoints answer 409 while Meshtastic is in use. See docs/API.md.

Not yet tested on real Meshtastic radios. It was tested against meshtasticd 2.7.26 and MeshHome's simulator. Direct messages between radios were tested with the simulator only, because meshtasticd's simulated radio doesn't encrypt them.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.1.1

Choose a tag to compare

@roach0816 roach0816 released this 12 Nov 20:02

Clearer Meshtastic setup: how to scan the team channel's QR code on each phone, and a Radio connection page that explains itself while the HAT runs Meshtastic.

The team channel's QR code

  • How to scan it, per phone:
    • iPhone: scan it with the Camera app and tap the meshtastic.org link. The Meshtastic app has no scan button on iPhone.
    • Android: use the scan button in the Meshtastic app's Settings → Channels.
    • Either way, choose Replace (not Add), so it becomes the tracker's main channel.
  • Open in Meshtastic imports the channel on the phone that's showing MeshHome, where there's nothing to scan. Copy link copies the same link.

While the HAT runs Meshtastic

  • Settings → Radio connection: the Radio HAT on this Pi option reads "Running Meshtastic, for Meshtastic trackers". Choosing it no longer says "Set up the radio HAT first". It explains that the MeshCore radio stays on None, and links to Settings → Tracking, where Meshtastic trackers are set up.
  • The radio status says "The radio HAT runs Meshtastic, for Meshtastic trackers" instead of "No radio configured yet".

Collection gaps: with the radio mode set to None, no collection gap stays open, since nothing is being missed. A gap that was left "ongoing" closes when you update.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over.

MeshHome 1.1.0

Choose a tag to compare

@roach0816 roach0816 released this 12 Nov 18:28

Meshtastic trackers, such as a SenseCAP X1 running Meshtastic, can now be followed on the Tracking map alongside MeshCore trackers. The radio HAT can run Meshtastic to hear them.

Meshtastic trackers

  • Nothing is polled. Meshtastic trackers broadcast their position on their own (on a timer, or when they've moved with smart position), and MeshHome records every position it hears from a tracked node.
  • Course progress, pace, ETA and offline maps work the same as for MeshCore trackers.
  • The details shown:
    • The battery level the tracker reports.
    • Its temperature and pressure.
    • How the packet arrived: "via Car Router (2 hops) · SNR 6 dB", or "heard directly".
  • Add one from Tracking → Add a tracker → Meshtastic node, which lists the nodes the Meshtastic radio has heard, trackers first.
  • Ask now sends a position request. A Meshtastic tracker answers at most every 3 minutes.
  • After a reconnect, positions heard while MeshHome was disconnected are filled in from the node's own list.

Where they're heard (Settings → Tracking → Meshtastic trackers)

  • The radio HAT on this Pi, running Meshtastic (below).
  • A Meshtastic node on the network, through its API on TCP port 4403 (for example an ESP32 node with Wi-Fi).
  • Simulated: two trackers running your first visible course, plus a router at the car, to rehearse at home.

A private team channel

  • Why: on Meshtastic's default public channel, positions are sent blurred to about 1.5 km, and anyone nearby can follow them.
  • Setting it up: Set up a team channel gives the Meshtastic radio a private channel with a random key and positions at full precision, plus your region and modem preset.
  • The QR code: scan it with the Meshtastic app into each tracker and relay (Channels → scan → Replace). That also matches their radio settings to MeshHome's. The code holds the channel key, so keep it to your team.

The radio HAT can run Meshtastic

  • Switching: Settings → Radio connection → Radio HAT on this Pi → Switch to Meshtastic, choosing the RAK module (RAK13300 or RAK13302) and its IO slot. Or run sudo meshhome radio-hat setup --firmware meshtastic.
  • What it does:
    • Installs meshtasticd (Meshtastic for Linux) with apt from Meshtastic's own repository, after checking the repository key's fingerprint.
    • Keeps meshtasticd reachable from the Pi only.
  • While it runs Meshtastic, the HAT isn't MeshHome's MeshCore radio: that radio mode is turned off and remembered.
  • Switch to MeshCore brings ZephCore back with its identity, contacts and channels, and Meshtastic's channels are kept for next time.

API change

  • For Meshtastic trackers, contact_id and public_key are null, and the new source field reads meshtastic with a node_id. Scripts that read trackers through an API key should handle that.
  • New endpoints: /api/settings/meshtastic and /api/tracking/meshtastic/…. See docs/API.md.

Licences: MeshHome stays MIT. It talks to meshtasticd (GPL-3.0, installed separately) over its network API, using a reader written for MeshHome.

Not yet tested on real Meshtastic hardware. This was tested against meshtasticd 2.7.26 and in Debian 13 containers.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over, and nothing changes until you choose Meshtastic.

MeshHome 1.0.20

Choose a tag to compare

@roach0816 roach0816 released this 12 Nov 03:49

More on each tracker: how far along the course, pace and ETA, battery percentage, which repeater its answer came through, and its sensor readings, in metric or imperial units.

Course progress

  • Each runner's position is placed on their course: "Red Loop · 3.2 of 6.5 mi · 49%", or "Off Red Loop (400 ft away)" when they're more than about 120 m (400 ft) from the line, or finished.
  • Choose the loop in the tracker's Edit dialog. At a handoff, change the label and the course together. Automatic uses the course the tracker's position is on, which can be wrong where loops share trail.
  • Start and finish: where a loop's start and finish meet, and on out-and-back sections, MeshHome follows on from the previous position, so a runner coming in reads 99%, not 1%.

Pace and ETA

  • "9:35 /mi · ETA 3:42 PM (20 min)" (or per km): pace is the distance covered along the course over the last ten minutes, and the ETA assumes that pace for the rest of the loop. A good cue for waking the next runner.
  • Without a course: pace uses the straight-line distance between fixes, which reads slower on twisty trail. A runner who has stopped shows Not moving.

Battery percentage

  • The battery shows as an approximate percentage worked out from the voltage, in red at 20% or below. The voltage is still shown in the list.
  • High-voltage batteries: a tracker that has ever reported over 4.25 V is taken to have a high-voltage cell (full at about 4.35 V), like the SenseCAP MeshTracker X1's. It reads lower for the same voltage, so it doesn't sit at 100% for the first hour.

Which repeater it came through

  • "via Car Repeater · SNR 4.8 dB", or "heard directly", read from the radio's log of the packets it hears. Tap the tracker on the map for the whole route and RSSI. This shows where your coverage hands over between repeaters along the course.
  • Repeaters are named after your contacts whose key starts the same way.

Sensors

  • If the tracker shares its environment readings (MeshCore app → Telemetry → Environment), its temperature, barometric pressure and altitude from the barometer are shown too.
  • The temperature is measured inside the tracker, so in a pocket it mostly reads the runner's body heat.
  • Trackers whose firmware doesn't report a barometer simply don't show these.

Metric or imperial

  • One Units setting for all of MeshHome: metric (km, m, °C, hPa) or imperial (mi, ft, °F, inHg). It covers the Tracking page (course distances, pace per km or per mile, off-course distance, altitude, temperature, pressure), the radio's sensors and the /env bot reply.
  • Where to change it: in Settings → Tracking, or where it already was, with the radio's sensors. Changing it in one place changes both.
  • Your existing choice carries over: if you'd already chosen imperial for the sensors, Tracking uses imperial too. The API always uses metric.

Rehearse at home: the simulated radio's tracker now reports a slowly draining high-voltage battery and sensor readings. Its answers come through the simulated repeaters.

Not yet tested on a real tracker. Please report how progress, the repeater shown and the sensor readings look on your X1s.

Upgrading: update from Settings → Software, or run sudo meshhome update on the Pi. For containers, pull the new image (Kubernetes/Fleet picks up the pinned image automatically). Your data and settings carry over, and the new columns are added automatically.