Changelog

Release notes for VoxDMR. Each release page on GitHub has the full commit list and the signed binaries; this is the human summary.

v0.12.0

DesktopMobile

Released June 2026. Desktop (now incl. macOS) + Android.

Two headline changes: VoxDMR now rides out connection drops on its own, and macOS gets its first official build.

  • Auto-reconnect. When the link drops unexpectedly — a timeout, a server-side error, or a network change (Wi-Fi ↔ cellular) — VoxDMR reconnects on its own with exponential backoff (~2 s growing to a ~60 s cap, jittered, unlimited retries) and re-subscribes to your last talkgroup, so you land back where you were. It stops only on a deliberate disconnect or an authentication failure. On by default; toggle it in Settings → Connection on desktop or Settings on Android. Shared by both frontends. See Auto-reconnect.
  • macOS desktop build. macOS is now a release target — a native Apple Silicon (arm64) .app shipped in a .dmg, with an in-app microphone-permission prompt on first launch and smooth CoreAudio receive playback. This clears the “macOS isn’t a release target yet” limitation noted back in v0.7.0. Intel Macs aren’t built yet.

config.toml (desktop) and the Android settings gain an auto_reconnect flag, defaulting to on; older configs without it pick up that default. No other config, paths, or protocol semantics changed.

v0.10.0

DesktopMobile

Released May 2026. Desktop + Android.

The headline change: VoxDMR is no longer BrandMeister-only, and no longer single-server. You can configure as many profiles as you want, each pinned to its own network and credentials, and switch between them with a tap.

  • Homebrew (MMDVM_HBP) protocol alongside the existing Rewind protocol. That brings TGIF Network, FreeDMR, ADN.systems, and any other MMDVM-based DMR network into VoxDMR. The TX/RX path is the same — Homebrew traffic encodes audio with the same AMBE+2 vocoder and shows up in the same call card.
  • Curated Homebrew server list. When you pick the Others protocol, the server picker offers TGIF Network, FreeDMR United Kingdom, and ADN Portugal (2681) out of the box, plus a Custom server… option for anything not listed. The list ships in the binary; PR additions on GitHub.
  • Hash format toggle for Custom Homebrew servers — Raw (the HBlink/MMDVMHost convention used by every major network) or Hex ASCII, for the rare network that wants the old format. Curated entries default to Raw.
  • Server Profiles. Settings → Connection grew a PROFILES card. Each profile bundles a label, DMR ID, callsign, protocol, server, password, and its own talkgroup favorites + aliases. Switching profiles disconnects, swaps the whole config, and reconnects.
  • Per-profile talkgroup aliases. Right-click a TG (or long-press on Android) to give it a custom name that overrides the database. The alias is scoped to the active profile — your BrandMeister and TGIF profiles can label 91 differently. Falls back to the official name when no alias is set.
  • Sectioned TG picker. The picker on the main window now groups results into DMR ID (when you type a numeric ID), Favourites, and Results. Headers appear only when they have rows. The DMR ID section exposes both Use as talkgroup and Use as private call affordances so custom IDs become first-class.
  • Android app. The whole protocol/profile stack ships in the Android client too — same Rust core, Material 3 Flutter UI, bottom-sheet server pickers, long-press to rename. Available now on Google Play.
  • Localisation. Settings, the setup wizard, modals, toasts, and main-view labels are extracted into a string catalogue. English and Portuguese (Portugal) are available now; the language picker lives in Settings → Interface on desktop and Settings → About on Android.

Config migration: existing config.toml files get the active connection promoted to a single profile labelled “Default” with the Rewind protocol — your DMR ID, password, master, favorites, and aliases are preserved untouched.

v0.9.0

Desktop

Released May 2026.

Adds in-app auto-update. VoxDMR now checks for new releases on launch and on demand, and can install them without leaving the app — no separate download trip, no installer to find.

  • Prompt on launch. When a newer build is published, VoxDMR opens a modal with the version, a What’s new link, and three choices: Update now, Skip this version (silenced forever), or Remind me later (24-hour cooldown).
  • Manual check. Settings → About has a Check for updates button. The result — Checking…, You’re on the latest, Update available, or an error — appears inline below the button so you don’t have to glance away to read it.
  • Verified, atomic install. The binary is downloaded with progress + cancel, SHA-256 verified against the release’s SHA256SUMS file, and atomically swapped into place — Windows-safe, even with the .exe running. A mismatched checksum aborts and never overwrites the existing build.
  • Clear post-install state. When the install finishes you get a one-click Restart now, or you can keep working and pick up the new build on next launch. If you dismissed the modal mid-flow, the footer status line confirms the install in green so the work isn’t silent.
  • Footer-as-toast. Transient update messages share the right-hand status slot in the footer for ~5 seconds instead of pushing the rest of the UI down.

Stable channel only — pre-releases are excluded. The check hits a fixed public repo (jcalado/voxdmr-site); no telemetry, no analytics, no registration. config.toml gains an [updates] section that records skipped versions and the Remind me later timestamp; older configs without it pick up the documented defaults.

v0.8.0

Desktop

Released May 2026.

Adds a Time-Out Timer (TOT) so a stuck PTT or a long-winded transmission can’t hold a talkgroup indefinitely — standard radio behavior, finally available client-side.

  • Four modes: Off, Warn only, Warn then cutoff (default), Hard cutoff. Configurable in Settings → Interface.
  • Visual + audible warning. The TX timer under the PTT button counts down (gray → amber → red); a short beep plays through your local audio output device at the warning threshold and again on cutoff. Sidetone never goes to the network.
  • Sensible defaults. 180 s duration, 15 s warning lead, both warning surfaces on. Duration capped at 190 s — anything higher would be silently dropped by BrandMeister’s own forwarding cutoff.
  • Auto-release on cutoff. Reaching the duration triggers a clean PTT release for you; press fresh to keep talking.

Adjacent UX polish:

  • The top status card now flips to Transmitting → TG … in red while you’re keyed up, instead of staying on Idle.
  • The settings window opens 150 px taller and the right-hand panel scrolls when content overflows, so growing tabs (Interface in particular) stay usable on smaller screens.

config.toml gains a new [tot] section; older configs without it pick up the documented defaults. No env vars, paths, or protocol semantics changed.

v0.7.0 (first public release)

Desktop

Released April 2026. Linux x86_64 + Windows x86_64.

The first build of VoxDMR distributed as a downloadable binary. Everything before this was internal and not announced.

What ships:

  • BrandMeister DMR client with full RX and TX over the Rewind protocol. Authenticate with your DMR ID and a hotspot security password, subscribe to talkgroups, transmit and receive AMBE+2 voice.
  • Runtime firmware loading. The MD-380 firmware needed by the AMBE+2 vocoder is fetched from third-party sources on first launch (or supplied manually for offline machines). The binary itself contains zero firmware bytes, which is what makes shipping it possible. SHA-256 verified before being accepted.
  • Live talkgroup activity indicator. Favorites show a colored dot driven by BrandMeister’s last-heard WebSocket feed: green = active, amber = recent traffic with the speaker’s callsign, gray = idle. 30-second TTL.
  • Configurable PTT. Push-to-talk or toggle modes. Spacebar by default; rebind to almost any single key.
  • Audio meters. 24-segment LED-style TX and RX meters with peak-hold, color zones (green / yellow / red), and a CLIP latch indicator.
  • Talkgroup picker. Search the bundled BrandMeister database by name or ID, save favorites, drag-reorder, mark private-call destinations. Custom IDs accepted via Enter.
  • Live master directory. The connection settings pull the current list of BrandMeister masters at launch, with a baked-in fallback if the API is unreachable.
  • Cross-platform. Native Linux (ALSA / PipeWire / PulseAudio via cpal) and Windows (WASAPI). Single binary per platform; no installer, no system services.

Known limitations:

  • macOS isn’t a release target yet.
  • The Windows binary isn’t code-signed yet. SmartScreen will prompt on first launch. Click More info → Run anyway.
  • HTTPS downloads (firmware fetch, master directory) don’t honour system proxy settings. Use the manual-files path or run from an unproxied network.

See the release page for binaries and SHA-256 sums, and the installation guide to get going.

Previous
← Troubleshooting