MULTILARM v1.2026
v1.2026 · build 2026.04

MULTILARM

multilarm.com · v1.2026 · Free for personal use · Commercial licence required

Schedule audio.
Anywhere. Offline.

Automatic school-bell, Azan player and PA announcement software for Windows, Linux & Raspberry Pi. One binary rings school period bells, factory break & shift bells, plays the Adhan/Azan in a mosque and drives a multi-zone public-address system — fully offline, with optional live PA, cloud remote control, MQTT & webhook integration. Built in VB.NET / .NET 8, runs as a service.

Free forever for personal, non-commercial use — no account, no trial clock, no feature lock. Download it and run it at home today. A licence is required to use it in a business, school, mosque or other organisation.

Windows Linux / RPi Offline MP3 OGG FLAC WAV TTS
01 // Overview

What Multilarm actually does.

Multilarm fires the right audio at the exact second, every day of the year. Install it on Windows, Linux or a Raspberry Pi — works fully offline and survives a router outage that would brick any cloud-only scheduling appliance.

The same engine that rings the period bell in a school drops shift announcements on a factory floor, signals a nurse-call routine via MQTT, calls Fajr in a mosque, and shuffles a community radio overnight. One XML config. One binary. One audit trail.

People come to Multilarm looking for an automatic school bell system, an auto Azan / Adhan player for mosque speakers, a factory break-bell or buzzer system, a shift-change siren, a Raspberry Pi PA system, or IP paging and public-address scheduling software — and find that one download does all of it, free for personal use and licensable for any site.

Built and maintained by Ali Muhammad on multilarm.com. Free for personal, non-commercial use — any business, institutional or paid use requires a commercial licence. See pricing.

SchoolsFactoriesHospitalsCare homesHotelsTheatresMuseumsCommunity radioMosques
The Multilarm Box — a sealed, wall-mounted appliance with a Raspberry Pi running Multilarm inside, a nine-light diagnostic status panel showing Power, Network, Wi-Fi and Cloud online with the next Dhuhr bell scheduled for 13:00, and a labelled connector edge (USB-C power, Ethernet, USB bell relay, emergency button, mic in, speaker out, line out DAC and microSD).
The Multilarm Box — concept render of the sealed wall-mounted appliance (Model MB-1).
02 // Why Multilarm

What it does differently

Offline first

No internet required for scheduling. Optional cloud relay is purely additive. Polls every 5 s, exponential backoff, survives extended outages without missing an alarm.

Live PA, not just playback

Push-to-talk from a local mic, the LAN dashboard or your phone over the cloud relay. Sub-second latency, P1 priority interrupt, runs alongside scheduled audio without a separate mixer.

Runs on a Pi

A full institutional PA on a Raspberry Pi. One-command Linux install, systemd at boot, headless. Same single binary scales from a 1-zone mosque to a multi-zone hospital fleet via Hub mode.

03 // Capabilities

Beyond scheduled playback

Multilarm isn’t just a timed media player. The same binary handles live PA, integrates with your facility systems, broadcasts to peers, and produces an audit trail your compliance officer will actually read.

Live PA & Push-to-Talk

Local mic from the console, the LAN dashboard or a browser anywhere via cloud relay. Sub-second latency, P1 priority interrupt, alarms queue automatically while you’re live.

External Trigger API

POST to /api/trigger/fire from any system. Bearer tokens with per-token allowlists, rate limits, cooldowns. Fire alarms from a button, a sensor, a panic key.

MQTT bridge

Subscribe to your existing broker for nurse-call, BMS, fire-panel or facility integration. Publish heartbeat & per-event status back out. TLS, base-topic isolation, JSON payloads.

Outbound webhooks

HTTPS POST on every alarm.fired, ptt.start, device.offline. HMAC-SHA256 signing (GitHub-format), retry with backoff, per-endpoint queue.

Hub fleet mode

Multi-instance fan-out across one site or a regional estate. Discover peers via heartbeat, proxy controls, broadcast voice memos. Promote/step-down at runtime; no central server required.

Operators & alerts

Admin/operator role split, signed audit log per device, email alerts on offline / missed alarm / missing audio. SMTP relay, webhook backup, rising-edge deduplication.

Also bundled: Kokoro 82M neural TTS · Windows Service install · System tray companion · Web file manager · Multi-zone mirror output
01 // Tools

Web tools & cloud

All downloads
Get a cloud account Watch the video guide Documentation
04 // Try it

Quickstart — one command

Full instructions

Windows

# Run the MSI installer
Multilarm-x64.msi
# Then launch from Start menu
Download MSI →

Linux / Raspberry Pi

# One-command install
curl -sSLO multilarm.com/multilarm/binaries/multilarminstall.sh
sudo bash multilarminstall.sh
multilarm
Linux setup details →

Link it (optional)

# Device shows a short setup code
multilarm --claim
# Enter it at multilarm.com/setup
How linking works →

Multilarm runs fully offline — linking is only needed if you want to manage the device from anywhere. Either way there is no XML to hand-edit and no token to copy: pick a preset for your kind of site (mosque, school, care home or office), and the setup wizard fills in the rest.

05 // Get Multilarm

Pick your platform

All downloads

Windows MSI installerRecommended

All-in-one installer for 64-bit Windows. Installs the binary, miniaudio engine, default config, sample audio and resources. No additional dependencies.

Min OSWindows 10 (64-bit) Architecturex64 Dependencies.NET 8 (bundled) Size~38 MB

Linux & Raspberry PiRPi-ready

Easiest on a Raspberry Pi: flash the ready-made SD-card image (Multilarm pre-installed, first-boot wizard for Wi-Fi, audio, and linking to the cloud with a short setup code — no token to copy, no XML to edit). Or use the one-command install script on an existing system — it handles dependencies, audio driver setup and a systemd service.

Architecturesarm64, arm32 (x86_64 on request) Tested onRaspberry Pi OS, Debian, Ubuntu AudioALSA / PulseAudio / Bluetooth Servicesystemd template included
03 // Configure

Configurations & presets

All 97 settings
3
Platforms
71
Config settings
82M
On-device TTS
100%
Offline-capable
04 // Solutions

Schools, factories & facilities

All solutions
05 // Guides

How to set it up

All guides
06 // Field Notes

Real-world deployments

All posts
07 // Questions

Frequently asked

Is Multilarm free?
Free for personal, non-commercial use — private hobby projects, personal productivity, learning. Any use within a business, institution, or as part of a paid service requires a separate commercial licence. That includes mosques, schools, hospitals, factories, hotels and any other organisational deployment. See pricing for full licence, cloud and commissioning rates. Source code is proprietary — binaries only.
Do I need an internet connection to run Multilarm?
No. Multilarm runs fully offline once installed. The only optional online feature is the Cloud Relay dashboard, which lets you control your instance from anywhere over HTTPS — the scheduler itself never needs the network to fire alarms.
What text-to-speech voices are available?
Multilarm ships Kokoro 82M, a neural TTS engine licensed Apache 2.0. Bundled in the EXE — works offline immediately on first launch, no separate install. British & American voices (e.g. bm_george, bf_emma, am_michael, af_heart), configurable speed 0.2–3.0×, sentence-level chunking. Cached WAV files generated on startup so the first announcement is instant.
Can I use my own Adhan recordings, bell sounds or PA jingles?
Yes. The bundled sample audio is all free-licensed (public-domain and Creative Commons recordings from Wikimedia Commons, with per-file attribution in CREDITS.txt; bleep tones and TTS words are original Multilarm recordings) — fine as a starting point, and most sites swap in their own licensed recordings (their preferred Qari, school chime, PA jingles), which Multilarm plays as-is. Drop any .mp3, .wav, .ogg, .flac, .aif or .aiff into the configured folder. Multilarm picks a random file per trigger by default; folder-specific shuffles and time-range schedules are configurable. Upload via the web file manager (16 MB max, magic-byte validated). Streaming sources (Shoutcast/Icecast) supported too.
Can it integrate with our nurse-call / BMS / external systems?
Three integration paths: MQTT bridge (subscribe to your broker, fire triggers from nurse-call or BMS messages, publish status/events); External Trigger HTTP API (Bearer-token POST to /api/trigger/fire with per-token rate limits, allowlists and cooldowns); and outbound webhooks on every alarm fire, PTT start/stop, device online/offline, with HMAC-SHA256 signing.
How do I set prayer times for my city?
Multilarm calculates them — you don’t need a timetable from anywhere else. Open the Salat Calendar Generator inside the Config Generator, enter your latitude and longitude (or tap Detect Location), choose the year, calculation method (MWL, ISNA, Egypt, Makkah/Umm al-Qura or Karachi), Asr madhab and high-latitude rule, and it writes a full year of Fajr, Sunrise, Dhuhr, Asr, Maghrib and Isha straight into the config — offline, with nothing fetched from the internet. Prebuilt configs also ship for UK Sheffield and Bangladesh, and you can paste your masjid’s own printed timetable instead if it differs.
Will it run on a Raspberry Pi Zero / Zero 2 W?
RPi Zero 2 W (ARM64) — yes, runs comfortably. Original Pi Zero (ARMv6) — not supported by .NET 8. RPi 3 and newer all work. Bluetooth and USB audio HATs documented in the install scripts. Multi-zone fan-out via MirrorDevices (one decode, shared clock across N output devices).
Is the Cloud Dashboard required?
Optional. The scheduler runs entirely offline. Cloud Relay adds: browser-based control from anywhere, multi-device fleet view, multi-operator permissioning (admin / operator roles), email alerts on offline / missed alarm / missing audio, audit trail, and browser push-to-talk to any device. Cloud is a subscription; LAN-only deployments don’t need it.
How is Multilarm different from cron + a media player?
Cron schedules a command. Multilarm schedules audio. You get a 3-priority mixer (P1 immediate, P2 on-quiet via miniaudio DSP tap, P3 queue), live PA with cloud streaming, MQTT & webhook integration, an external trigger API, Hub mode for fleet management, Kokoro neural TTS, a Windows Service + system tray companion, multi-zone mirror output, audit logging, email alerts, and an operator dashboard. None of that is cron.
Where do I report a bug or request a feature?
Email with the relevant snippet of Multilarm.error.log and your Multilarm.config.xml (redact device tokens). The fastest path to a fix is a reproducible example. See bug reporting in the docs.

Ready to schedule precision audio?

Free for personal use. Runs offline. Works on hardware you already own — from a $35 Pi to a multi-zone server.

↓ Get Multilarm for your platform