Comparison · Open-source radio station automation

Multilarm vs LibreTime: an honest comparison

The open-source radio automation suite: a calendar of shows, playlists and smart blocks, played out to Icecast for listeners on the internet.

Checked against published documentation on 2026-09-13

LibreTime is the right choice if you are running a radio station — a grid of shows, live presenters, and listeners on the internet. Multilarm is the right choice if the audio is for a building: bells, announcements, background music, priority and emergency handling on speakers you already own. Multilarm has no Icecast source output and no remote DJ, and publishes both as absent rather than as roadmap.

This page is written by the makers of one of the two products, so the only version worth publishing is one that is specific about where the other one wins. Every section below names a buyer who should not choose Multilarm.

What LibreTime is

LibreTime is a web-based radio automation system, the maintained community fork of Airtime. A calendar holds shows, shows hold playlists and smart blocks, and a playout chain streams the result to Icecast, with live sources able to take over for a presenter. It is self-hosted, free, and licensed under the AGPL.

It is on this list because scheduled background music in a building and scheduled programming on a station look like the same problem right up until you get close to either one.

Where LibreTime is better

Public internet streaming. LibreTime exists to put a programme on an Icecast server for listeners anywhere, with listener statistics to match. Multilarm publishes this as an absent feature: its audio engine encodes WAV and FLAC only, so there is no lossy encoder to feed a streaming server with.
Live presenters. A remote DJ can connect a source client and take the output over, with the automation picking up again afterwards. Multilarm has microphone passthrough for somebody standing at the machine and nothing at all for a presenter somewhere else.
A show-oriented calendar. Programming is organised as shows with hosts, repeats and rebroadcasts, which is how a station actually thinks about its week; Multilarm thinks in schedules, routines and rotations.
It is free and open source under the AGPL, with the whole codebase available to read, change and keep running whatever happens to anybody who maintains it.
A community of station operators who have already solved station problems, and interface translations into many languages.

Where Multilarm is better

It is built for a building rather than a station: bells, timetables, term dates, shift patterns and an exam week that is not the same as an ordinary one.
Priority mixing and ducking. An announcement interrupts the music at the right level and the music comes back, with an emergency path that outranks everything else on the speakers.
Emergency handling with a physical button, a relay output and an audit ledger — alongside a published statement that it is not fire-alarm equipment and does not replace one.
It runs offline, with no inbound ports open, on a Raspberry Pi or an ordinary PC, and needs no Docker host for somebody to look after.
Offline text to speech, so a typed sentence becomes a spoken announcement with no internet connection and no recording session. English only, and published as English only.
Synchronised playback across separate machines, with a probe that measures the accuracy on your own network rather than asking you to take a number on trust.
Vendor support, signed updates, a diagnostics battery and a configuration tool aimed at somebody who is not a developer.

Side by side

Cells reading “Not published” mean we could not find the answer in Open source (LibreTime)’s own documentation on 2026-09-13. That is not the same as “no”, and this table will not pretend it is.

Capability	LibreTime	Multilarm
Streaming to public internet listeners	Yes — Icecast output with listener statistics	Not built — the engine encodes WAV and FLAC only, and features.json publishes this as absent
Live or remote presenter takeover	Yes — live sources can take the output over	Live microphone at the machine only
Show calendar with repeats and rebroadcasts	Yes	Schedules, routines and a weighted rotation rather than a show grid
Smart playlists and rotation separation rules	Yes — smart blocks	Yes — smart playlists, weighted rotation, artist and track separation, and a dead-air guard
Play log and reporting-return exports	Yes	Yes — monthly CSV, with PRS, PPL and SoundExchange shapes
Podcast publishing	Yes	Yes — scheduled recordings offered as an RSS feed
Scheduled bells and a school or shift timetable	Out of scope by design — it is a station product	Yes
Announcements ducked over the programme	Out of scope by design	Yes
Offline text to speech	Out of scope by design	Yes — offline, English only
Emergency button, relay output and audit ledger	Out of scope by design	Yes — and it is not fire-alarm equipment
Listening on the local network with nothing to install	Via the public stream	Yes — an optional LAN endpoint serving uncompressed audio to a browser
Runs with no internet connection	Not published as a supported mode	Yes — offline is the normal case
Install shape	Docker on a Linux server	Windows installer, Debian package, Raspberry Pi image or a prebuilt appliance
Cost	Free and open source (AGPL)	Free for personal non-commercial use; perpetual licence otherwise
Support	Community	Vendor

Choose LibreTime if

You are running a radio station and the listeners are on the internet.
Remote presenters need to go live on air.
You think in shows and a programme grid rather than in bells, announcements and zones.
You want open source, and a Linux server running Docker is not a problem for you.

Choose Multilarm if

The audio is for a building: bells, announcements, prayer times or shift changes on speakers you already own.
Something has to interrupt the music at the right level, and an emergency has to outrank everything.
It must keep working with no internet connection and no inbound ports open.
Whoever operates it is not a developer, and there is nobody to look after a Docker host.

How to check any of this

Multilarm’s own capabilities are published as machine-readable data at features.json, including every feature it does not have and the ones it has deliberately declined to build. Measured performance figures, with the method for each, are on the measured data page. The canonical facts about licensing, offline operation and the fire-alarm boundary are at facts.json. If anything on this page is wrong, say so and it will be corrected.

FAQ

Can Multilarm stream to Icecast the way LibreTime does?

It cannot. The in-house audio engine encodes WAV and FLAC only, so there is no lossy encoder to feed a streaming server with, and that is published as an absent feature rather than as a roadmap promise. There is an optional LAN listen endpoint that serves uncompressed audio to a browser on the same network, which is a different thing at a different scale and is described that way everywhere.

Could a station run both?

Yes, and in a building that also broadcasts it is a reasonable answer: LibreTime for the programme going out to the internet, Multilarm for the bells, announcements, zones and emergency audio inside the building.

Does Multilarm handle music licensing?

It produces the paperwork — a play log with exports shaped for PRS, PPL and SoundExchange returns — and takes no view on whether a licence is needed, which one, or what it costs. That is between you and the licensing society.

Sources

Everything stated here about another product is taken from that vendor's own public documentation on the date shown, and is limited to what they publish about themselves. Products change; if something below is out of date or wrong, tell us and it will be corrected. Where we could not find a published answer the table says so rather than guessing.

Not sure this is the right purchase for your building? Ask a specific question — including whether you should be buying something else.