mrefd is the reflector software for M17, the newest of the common amateur digital voice modes. It rounds out the family of explainers alongside XLX and URF — but where those wrestle with bridging incompatible codecs, mrefd is refreshingly simple, and that simplicity is the whole story. M17 is a single, fully open mode, so an M17 reflector never has to transcode anything. It just connects M17 users together and passes their audio along.
mrefd is a server that acts as a central meeting point for M17 radio — the same basic idea as XLX or URF, but dedicated to one mode. M17 hotspots, repeaters, and other mrefd reflectors connect to it over the internet, and once connected, their users can all talk to one another. An incoming M17 voice stream from any one client is simply sent out to all the others. (M17 also has a data "packet" mode, and mrefd handles that too.)
Like XLX and URF, an mrefd reflector is divided into modules, labeled by letter from A to Z — up to 26 independent talk rooms. Everyone linked to Module C hears each other, and that conversation is completely separate from Module D. The operator decides how many to enable and what each one is for. A client links to whichever module it wants by name.
This is what sets an M17 reflector apart from the others, and it's worth dwelling on. The XLX and URF pages spend a lot of time on transcoding — hardware vocoders, codec gaps, the cost of bridging D-Star to everything else. None of that exists here.
Why it's so much simpler
M17 uses Codec2, a fully open-source voice codec, and every M17 user is on M17. There is no second codec to bridge to, so there is nothing to transcode — no AMBE chips, no vocoder dongles, no hardware to buy. The reflector's only job is to relay M17 to M17. That makes an M17 reflector dramatically cheaper and simpler to run than a transcoding XLX or URF reflector.
It also means there's no codec-licensing baggage. The proprietary AMBE and AMBE+2 codecs that make D-Star, DMR, and Fusion transcoding expensive simply aren't part of the M17 world. M17 was designed from the start to be open, and a reflector built for it inherits that openness.
Because there's no transcoding, the reflector's work is mechanical and lightweight. When a voice stream arrives from a client, mrefd re-addresses it to each of the other clients on that module — rewriting the destination callsign in the packet and recalculating its checksum — then sends it on its way. No audio is decoded or re-encoded; the original M17 stream is just relabeled and forwarded. That lightness is why an M17 reflector can run comfortably on very modest hardware.
mrefd reflectors can be interlinked into peer groups, extending a module across multiple servers. A few rules shape how this works: an interlink has to be configured on both sides, and within a group every reflector links to every other reflector (a full mesh), which keeps latency between any two users as low as possible. To prevent audio from looping endlessly around the group, a voice stream is marked when it's passed from one reflector to another, and the receiving reflector knows not to pass it on again — a one-hop rule borrowed from the XLX design. One limitation: you can link the same module across reflectors, but you cannot cross-link different modules (say, Module A on one reflector to Module B on another).
mrefd is the specialist of the family: one mode, one open codec, no transcoding, no special hardware. Where XLX and URF earn their complexity by bridging incompatible modes, an M17 reflector sidesteps the entire problem because M17 was built open from the start. The result is a reflector that's cheap to run, light on hardware, and simple to understand — connect M17 clients, pass the audio along, done.
mrefd is written and maintained by Thomas A. Early (N7TAE), the same developer behind the improved XLX fork and the URF reflector. In fact, much of mrefd's code is originally based on the groundbreaking XLXD work, and the copyrights on those source files reflect that shared heritage. It's published as open source under the GPL, part of the broader M17 Project — a community effort to build a completely open digital voice ecosystem, free of proprietary codecs and licensing. mrefd is the reflector piece of that vision.
A noncommercial hobby reference compiled by N6JET, gathered from public sources and shared freely for anyone interested in amateur digital voice.