AudioBridger — send any audio between any computers, with no setup

What’s new in 1.2.0

A working session’s worth of upgrades to the Node Graph editor and the routing layer beneath it.

Chain presets & junction boxes. Save a chain as a reusable preset and drop it onto any wire as a single card. Junction nodes mark shared patch points so fan-out topology survives an FX delete — rewire later without losing the structure.
Multi-select & multi-drag for FX. Marquee or shift-click effects to move them as a group. Selections survive the live status poll instead of vanishing on the next tick.
Quiet startup window. The dashboard waits for mDNS discovery to settle before drawing any wires, so you stop seeing wires snap between placeholder cards while senders and receivers are still announcing themselves. A small pill in the corner tells you what’s happening.
Bullet-proof chain integrity. Routes and effect chains can no longer get into “zombie” states where audio silently stops because of a broken DSP path. The hub refuses to persist incomplete chains; the licence cap only counts senders whose chain has a complete sender → receiver path.
Receiver VU meter reflects post-DSP output. Used to show pre-DSP packet peaks — two receivers subscribed to the same source mirrored each other’s meters even when one had a broken chain. Fixed: meters now show what’s actually reaching the speaker.
Ten built-in effects. Gain, pan, hard limiter, compressor, 5-band parametric EQ, noise gate, de-esser, delay, reverb, saturator — every one fully wireable into the node graph.

Read the full changelog →

What it does

A short list, because that's what AudioBridger is — short and to the point.

⚡

Low latency

Designed for live monitoring, not file transfer. Audio flows in tens of milliseconds, not seconds.

🔌

Zero setup

Open it on two computers and they find each other automatically over your network.

🕸️

Node-based patchbay

Route your whole setup on an infinite canvas, like a modular patchbay. Every device is a node you can drag, box-select and move in groups, rename, and colour-code by role. Drag from a sender to a receiver to wire them — the connection goes live instantly. Pan, zoom and auto-arrange to keep even a big rig readable.

👥

True many-to-many routing

Wire any input to any output across your network — any number of senders to any number of receivers, in any combination. A receiver can listen to several senders at once and mix them together; a single sender can feed up to 10 receivers. Edit the whole web live: connections take effect the moment you draw them. Great for headphone mixes, shared monitoring, or combining sources.

🔊

Every device, one launch

Run it once on a machine and it brings up every audio device on that computer — each input as its own sender, each output as its own receiver. No more launching a copy per device or juggling separate folders. Pick a subset with a flag if you'd rather.

🎛️

Per-source mixer

When a receiver is fed by multiple senders, the hub gives you a channel-strip mixer for it: independent gain, pan and mute on each incoming source, plus an overall output fader — all with live VU meters and dB markings.

🎚️

Per-source effect chains

Each sender feeding a receiver gets its own DSP chain — gain, pan, hard limiter, compressor, 5‑band parametric EQ, noise gate, de‑esser, delay, reverb and saturator. Compressor has a draggable transfer-curve graph; EQ has a FabFilter-style frequency response with draggable band points. Built-in presets plus “Save as preset” for your own. Edits are live: drag a slider in the dashboard and the receiver applies the change on the next audio buffer (~1 ms). Chains persist in the hub config and replicate across HA hubs.

🕸️

Node-graph editor

The wires are the truth. No grid mixer, no routing matrix, no checkbox table to puzzle out. If you can see a wire, it’s connected; if you can’t, it isn’t. DAWs make you reason about signal flow in spreadsheets — AudioBridger just shows you the signal flow.

Build effect chains by dragging wires between ports on the canvas — sender out, FX in/out, receiver in. Drop an FX card anywhere first and wire it later. Fan one FX’s output to multiple receivers and edits stay in lock-step across every mirror. Cut a wire and it stays cut: no auto-reconnect, no surprises. Same mental model as Blender, DaVinci Resolve or Houdini — applied to audio.

Save any chain as a preset and drag it onto another wire as a single card — expand to edit the inner FX, or treat it as a black box. Drop a junction node to mark a shared patch point; deleting an FX that fed multiple downstream receivers leaves the junction in its place so audio keeps flowing while you rewire.

🎛️

Sample-rate conversion

Mismatched sample rates used to be silently rejected; now they get a linear-interp resampler on the fly so audio actually flows. A yellow hazard chip on the mixer row tells you which source is being converted (and to what rate), so you can fix it at source for best quality.

🖥️

Cross-platform

Mac and Windows talk to each other freely — a Mac can stream to a Windows machine and back, in either direction.

🎚️

Pick any input

Mics and line-ins work out of the box. To send system audio or a specific app, both macOS and Windows need a small free helper (BlackHole, Loopback, VB-Cable, VoiceMeeter) that exposes the audio as a virtual input — once that's installed, AudioBridger picks it like any other device. See the helpers we recommend.

📊

Built-in dashboard

An optional web page shows every active sender and receiver, grouped by computer, with live VU meters and connection status. A List view, a drag-to-wire routing graph, the node-based patchbay, and a per-receiver mixer — route audio, mute, set gain/pan, and pick output devices remotely from any browser on the network.

🛡️

Redundant hubs

Run a second hub for high availability. The two stay in sync and agree on which one is in charge; if the active hub goes down, the standby takes over automatically with no lost routes. Your whole layout — connections and node positions — is mirrored across both, so either one shows the same picture. You can also promote a standby to primary instantly from the dashboard menu. Because audio is peer-to-peer, hub failover or reconnect never interrupts a stream.

🌐

Listen in a browser

Each sender also publishes a low-bitrate AAC stream over HTTP. The hub dashboard shows a play button on every sender that has it enabled, so you can monitor any source from any browser on the network — no receiver needed. The same URL works in a phone, VLC, or OBS. Requires FFmpeg installed on the sender; if it's missing, the AAC stream simply turns off and the rest keeps working.

🎵

Studio-quality audio

Uncompressed 16-bit PCM, up to 48 kHz stereo, with no codec artefacts on the main path. Uses about 1.5 Mbps per receiver — easy on any modern Wi-Fi or wired LAN, and small enough that a dozen streams fit comfortably. Senders at different sample rates mix into the same receiver: the mixer auto-resamples on the fly and shows a warning chip on any source that’s being converted, so you can match the rates at source for best quality whenever you want.

🔄

Auto-reconnect

Network blip? AudioBridger picks back up where it left off — no restart, no fiddling. Because audio flows peer-to-peer between sender and receiver, the hub can go away and come back without ever interrupting the stream. Receivers keep playing from their last known sender address until the hub delivers updated routing.

🪶

Lightweight

No daemons, no system extensions, no driver installs. Just a small app that runs when you want it.

Use Cases

Anywhere audio lives on one machine but needs to come out of another.

iOS & mobile development

Building an audio app in Xcode or Android Studio? Pipe the simulator's output to your real studio monitors instead of your laptop speakers. Hear what the app actually sounds like — bass, stereo image, the lot.

Virtual machines & remote desktops

Run a Windows VM on a headless server and need to hear what it's playing? AudioBridger sends the VM's sound to any computer on your network. Same trick works for remote desktop sessions, sandboxed apps, and dev containers.

Self-hosted AI & home-lab boxes

Running something like Ollama, OpenClaw, a media server, or a home-lab box that produces audio — alerts, voice replies, transcription playback — on a machine tucked away in a cupboard? AudioBridger sends that audio to wherever you actually are. Headless boxes stop being silent.

Podcasts & remote interviews

Bring a guest's mic into your recording rig from another room without re-routing audio interfaces. Each sender publishes a stream URL too, so producers can monitor on headphones from anywhere on the network.

Classrooms & group listening

Stream a single source — an instrument, a talk, a video — to every laptop in the room. Each student gets their own volume control, no shared speakers, no Bluetooth pairing.

Listen on a phone or laptop browser

Every sender can also publish a low-bitrate AAC stream over HTTP, so anyone on the network can monitor it in a phone, a browser, VLC, or OBS — no AudioBridger install needed. The hub dashboard adds an inline play button for each AAC-enabled sender, useful for engineers checking what's flowing without touching the source machine.

Studio & live monitoring

Send the mix from your studio computer to a tablet or laptop in the live room. Talent hears themselves with no cables run across the building. Multiple performers can listen to the same feed at once.

PA & venue distribution

Send a single audio source to multiple amplifiers around a building or venue — foyer, dressing rooms, satellite spaces. Anything with a network drop and a small computer becomes a listening point.

How it works

Three small apps. Two of them do the work; the third is optional and just gives you a dashboard.

🎙️

Sender Picks a sound source and broadcasts it.

→

📡

Network Your normal Wi-Fi or wired LAN.

→

🔊

Receiver Plays it through your chosen speakers or headphones.

Node view. See every sender and receiver on the network, grouped by machine. Drag nodes around to tidy up your layout, or box-select a group and move them together. Add effects and connect routes in a visual way, much easier to visualise than numbers

Graph view. Senders on the left, receivers on the right. Drag a wire from a sender's port to a receiver to send audio there. Grab a wire's endpoint to move it onto a different node, or drop it on empty space (or click the X on a receiver) to disconnect. Pending edits show as blue dashed wires until you hit Save.

List view. One row per sender and receiver, with live VU meters, packet counters, and connection status. Each sender row has an input-device picker; each receiver row has a mute toggle and an output-device picker, so you can change what a remote machine plays through — all from the browser. Senders with the AAC stream enabled get an inline play button too. Listen to the streams in the browser if it's avaialble too.

Sender

Create a stream source

Use any available device to create as your output stream and make it ready for broadcasting to the network.

Receiver

Capture system

The receiver finds the sender automatically over the network and plays it through any output device you choose.

The hub is control-plane only

Audio always travels directly between sender and receiver over peer-to-peer UDP — it never passes through the hub. The hub is the control plane: it manages routing intent, mute/gain/pan, the dashboard, and config persistence. If the hub goes offline, audio between already-connected senders and receivers keeps flowing uninterrupted. Only routing changes, mixer adjustments, and the dashboard itself require the hub to be reachable.

Not AES67 — on purpose

AES67 (used by systems like Dante, Livewire and Ravenna) is designed for high dependency broadcast environments. It typically requires:

Managed network switches with QoS
A PTP grandmaster clock
Per-node licensing

That makes sense in a network operations centre — but not in a project studio, classroom, home lab, developer virtual machines or home office.

AudioBridger is intentionally simpler:

Runs on your existing network (Wi-Fi included)
No managed switches required
No PTP or time clock
No per-seat fees

If you need AES67-grade interoperability with an existing Dante or Ravenna setup, then AudioBridger isn’t the right tool. If you just need to get audio from one machine to another without the overhead, then maybe it is.

Using AudioBridger

The two-app headless flow takes about thirty seconds. Add the optional gui hub view when you want a dashboard or remote routing - it's so simple to use.

Step 1

Run the sender

Open ab_sender on the machine whose audio you want to share. Pick an input device from the list when it asks — that can be a real microphone, a line-in, or a virtual cable (BlackHole, VB-Cable, VoiceMeeter) for system or app audio.

Step 2

Run the receiver

Open ab_receiver on the machine you want to play audio on. It finds the sender automatically over the LAN and connects within a second. Pick the output device — speakers, headphones, an audio interface — and you're streaming.

Step 3 (Recommended)

Run the hub

Open ab_hub on any machine on the network, then point a browser at http://<hub-ip>:4466. Every sender and receiver finds the hub automatically and appears grouped by computer. You get a live dashboard plus full remote control: drag wires to route audio, mix multiple sources per receiver, mute, set gain and pan, and swap output devices from anywhere.

In the hub dashboard

List view — senders and receivers grouped under each computer in collapsible sections, with live VU meters, packet counters, an input device picker for each sender, mute buttons, and an output device picker for each receiver. Senders with AAC enabled also get an inline play button so you can listen in the browser.
Graph view — inputs on the left, outputs on the right, grouped by machine, each node showing a live meter. Drag from a port to wire a connection; a receiver can take several sources at once. Grab a wire's endpoint to move it, click a wire and press Delete to remove it, or drop on empty space to disconnect.
Nodes view — an infinite-canvas patchbay. Every device is a card you can drag, box-select and move in groups, rename, and mute, colour-coded by role and labelled with its machine. Pan, zoom and auto-arrange to keep a big rig readable; drag from a sender to a receiver to wire it and the connection goes live instantly. Positions are shared across browsers and mirrored across redundant hubs.
Receiver Mixer — a channel strip per receiver: an output fader with dB markings and a folding row for each incoming source with its own gain, pan and mute, all live-metered.
Save Changes — pending route edits show as highlighted wires. Click Save and the hub immediately retargets the affected receivers; audio follows within a second.
Live meters — VU bars across every view update in real time over a WebSocket, so the dashboard tracks the signal without constant polling.
Hamburger menu — one place for admin actions: export the routing config to JSON or a full backup (config + identities + licence), import a saved config, take over the leadership of an HA pair, clear stale endpoints, copy the dashboard URL, manage the licence, and set an optional admin password that gates routing changes (read-only views stay open).

Choosing devices

By default the sender captures every input device and the receiver drives every output — one launch handles the whole machine.
--device "Name" or --device 0,2 to limit it to specific devices (by name fragment or index).
--device pick (sender) for the classic interactive single-device menu.
Each device gets a stable identity tied to its name, so saved hub routes stick to the right device across restarts.

Terminal controls

Both apps show a compact console listing every device they're running, each with a live L/R meter and status.
m (receiver) to mute / unmute all outputs locally. The hub can also mute per receiver remotely.
f (sender) — retry loading FFmpeg if you installed it after starting, enabling the AAC stream without a restart.
q to quit.

When something doesn't connect

Both machines on the same Wi-Fi or LAN. Mixing networks (one on guest Wi-Fi, one on Ethernet) won't let mDNS find them.
Allow the firewall prompt on first run. macOS and Windows both pop one up; click "Allow" for both ab_sender and ab_receiver.
If discovery is blocked, give the receiver the sender's IP directly: ab_receiver 192.168.1.50.
Same machine for sender and receiver works fine — useful for testing — but you'll need a virtual cable so the output of one doesn't loop back into the input of the other.

Licensing & activation

AudioBridger is honest software. You can run it for 30 minutes at a time without a licence — long enough to be sure it works for you. Buy a code, paste it into the app, and it runs for as long as you need.

Tiers

Trial — no licence. The sender and the hub each run for 30 minutes per launch and then quit cleanly. Receivers are always free, with or without a licence.
Personal — one code covers your whole cluster: up to 4 senders and unlimited receivers. Designed for a single studio, classroom, or home office.
Pro — same one-code-per-cluster model, but with unlimited senders. For venues, broadcast facilities, and multi-room installations.

Licences are perpetual — once activated, the version you bought keeps working forever. There's no subscription, no monthly fee, no per-seat charge. A licence is locked to a major version (v1, v2, …); future major releases may be a paid upgrade.

The full terms are in the End User Licence Agreement, which also ships in every download.

How activation works

Buy a licence from the store. You'll receive a licence code by email.
From the hub web UI: open http://<hub-ip>:4466 in a browser, click the red Activate button at the top, paste your code, and hit Activate. The trial timer cancels immediately and the hub keeps running.
From the command line: activate either the hub or a sender with ./ab_hub --activate <CODE> or ./ab_sender --activate <CODE>. A registration.lic file is written next to the binary and picked up on the next launch.
One code, one setup: a single licence covers your hub and senders. Activate on each machine with the same code and it's ready to go.
Automatic licence distribution: if the hub is already running with a valid licence, senders and receivers on the same LAN will fetch the licence file from it automatically at startup — no manual file copy to each machine needed.

Offline activation

On a machine with no internet, run ./ab_hub --offline-request <CODE>. It prints a short JSON payload.
Copy that to a connected machine and paste it into the offline form at bluetomatomedia.com/activation/audiobridger/offline.
You'll get a registration.lic file. Drop it next to the binary on the offline machine and re-run.

Moving to a new machine

Each code allows a small number of activations (typically 3). To free up a slot, deactivate the old machine first — see Manage licence in the hub web UI, or contact support.
If you've genuinely run out of activations and need more, just get in touch — we'd rather sort it out than have you blocked.

If activation fails

Check the code. Codes are case-insensitive and use an alphabet without confusable characters (no 0, 1, I, O), so a careful re-type usually fixes typos.
Check the network. The activation request goes to bluetomatomedia.com over HTTPS — most corporate proxies allow this, but the occasional firewall blocks it.
Already activated elsewhere? If the code has been used on its full quota of machines, you'll see an "activation count reached" message. Deactivate one of the older machines or contact support.

FAQ

Common questions before and after you start using it.

Does it need an internet connection?

No. AudioBridger runs entirely on your local network — the sender and receiver find each other over LAN and the audio never leaves your house, studio, or office.

The internet is only touched once, when you first activate a paid licence (a single HTTPS call to bluetomatomedia.com). After that the licence file sits next to the binary and never phones home again. Offline activation is supported too.

Will it work over Wi-Fi?

Yes. AudioBridger uses about 1.5 Mbps per receiver — well within any modern Wi-Fi link. A wired connection on at least one end gives you the most consistent latency, but Wi-Fi-to-Wi-Fi works fine for most setups. A muted source, or one nothing is listening to, stops sending almost entirely — so idle devices don't waste bandwidth.

The one thing that won't work is mixing networks — if the sender is on guest Wi-Fi and the receiver is on the main LAN, mDNS can't find them across the boundary. Same SSID or same wired LAN.

What's the latency?

Tens of milliseconds on a healthy LAN — designed for live monitoring, not playback. The exact figure depends on your buffer setting (default 20 ms), your network, and how chatty your Wi-Fi is at that moment.

For musicians and broadcast use it's well within "lip sync" territory. For pure listening it's instant.

How is this different from VBAN / VoiceMeeter?

VBAN is a genuinely great tool. It's been the go-to for free network audio for years, the protocol is public, and it pairs beautifully with VoiceMeeter on Windows. If you're already using it and it works for you, there's no reason to switch.

Where AudioBridger fits differently is the cross-platform story. VBAN's Mac story is mostly about receiving — VBAN Receptor on macOS and iOS lets you listen to a Windows source nicely. But for sending from a Mac, the picture is patchier: VoiceMeeter is Windows only, and you end up hunting for a third-party VBAN sender. AudioBridger ships first-class sender and receiver builds for both platforms, so every direction works the same: Mac to Windows, Windows to Mac, Mac to Mac, Windows to Windows.

The bigger difference is the optional hub. VBAN routing lives in VoiceMeeter on each Windows box — there's no central picture of the whole network and no way to reach into another machine. AudioBridger's hub is a web dashboard you open from any browser: a drag-and-drop node graph where you wire any input to any output across every machine, a per-receiver mixer with gain and pan on each source, and remote mute of any receiver — or any individual source feeding it — without touching that computer. You can also listen to any sender straight in the browser. If you don't need any of it, the standalone sender/receiver pair runs on its own.

One genuine VBAN advantage: native receiver apps for iOS and Android. AudioBridger doesn't ship mobile apps — though the in-browser AAC stream from each sender covers most phone listening cases without needing one.

How is this different from Dante, Loopback, or NDI?

Dante / AES67 are broadcast-grade audio-over-IP standards. They're excellent but expect managed switches with QoS, a PTP grandmaster clock, and per-node licensing. AudioBridger does none of that — it runs on your normal Wi-Fi.

Loopback (Rogue Amoeba) is a virtual audio cable for one machine. AudioBridger crosses machines. They actually pair well: Loopback grabs an app's output, AudioBridger sends it to another computer.

NDI is video-first with an audio sideband, aimed at broadcast workflows. AudioBridger is audio-only and tiny.

Can I send my computer's system audio (not just a mic)?

Yes, but both macOS and Windows hide system audio behind a virtual driver, so you need a small free helper:

macOS: BlackHole (free) or Loopback (paid). Windows: VoiceMeeter or VB-Cable (both free). Install one, then pick it as your input device when the sender starts.

How many computers can I run it on?

Receivers are unlimited and free, with or without a paid licence. Each sender can fan out to up to 10 receivers at once.

For senders: the trial allows one sender per launch, capped at 30 minutes. The Personal licence covers up to 4 senders in your network. The Pro licence is unlimited senders.

I bought a licence. How do I move it to a new machine?

Just activate with your code on the new machine. Each licence code allows a small number of activations (typically 3) to account for replacements; if you genuinely need more, email support and we'll sort it out.

The receiver can't find the sender. What's wrong?

Three things to check, in order:

1. Both machines on the same network. If one is on guest Wi-Fi and the other on the main LAN, mDNS won't cross. 2. The firewall prompt on first run was allowed (macOS and Windows both pop one up). 3. If mDNS is genuinely blocked, pass the sender's IP directly: ab_receiver 192.168.1.50.

Can I use AudioBridger commercially?

Yes — the Personal and Pro licences both cover commercial use within their tier limits (senders per cluster). See the licence agreement for the full terms.

What happens if the hub goes offline?

Audio keeps flowing. Because AudioBridger streams audio peer-to-peer directly between senders and receivers, the hub is not in the audio path at all. Already-connected pairs continue playing with no interruption for as long as the sender and receiver stay up — even if both hubs die.

What you lose while the hub is down: the ability to change routing, adjust mute/gain/pan, or see the dashboard. The moment the hub comes back, all endpoints re-register and resume normal operation. Receivers also hold their last known sender address while the hub is away, so there is no audio gap on reconnect.

Is the source code available?

AudioBridger is closed-source commercial software. The binaries are signed; the protocol is private. If you have a specific integration need, get in touch.

How we compare

AudioBridger sits in a specific spot: simpler than broadcast standards, more flexible than Windows-first tools - same but different...

Criterion	AudioBridger	VBAN / VoiceMeeter	Dante	NDI	Loopback (Rogue Amoeba)
Price	Trial free, paid per clusterperpetual	Freedonationware	Free for 2 channelspaid past that	Free SDK + free Tools	~$110 USDperpetual, mac only
macOS support (send + receive)	Native	Receive onlyvia VBAN Receptor	Via Dante Virtual Soundcard (paid)	Native	macOS onlythat's the whole point
Windows support (send + receive)	Native	Native	Via Dante Virtual Soundcard (paid)	Native	No
Mobile receivers (iOS / Android)	In-browser AAC streamno native app	VBAN Receptor on both	No	No (separate apps cost money)	No (local only)
Works on regular Wi-Fi	Yes	Yes	Possible, not recommendedwants managed switches	Possible, hungry on bandwidth	N/A (local only)
Needs managed switches / PTP clock	No	No	Yes, for serious use	No	No (no network at all)
Auto-discovery (zero config)	mDNS	Manual IP entry usual	Yes	Yes	N/A (local)
Multiple senders, multiple receivers	Yes, in any combination	Up to 8 streams	Up to 512 channels	Yes	Yes, but all on one Mac
Built-in mixer (EQ, compression, faders)	Yesgain, pan, hard limiter, RMS compressor, 5‑band parametric EQ, noise gate, de‑esser, delay, reverb, saturator — per source	YesVoiceMeeter is a full mixer	Via Dante Controller	No	Per-source volume; no EQ/comp (use Audio Hijack)
Node-graph effects editor	YesBlender / Resolve-style, fan-out, cut-stays-cut	No	No	No	Wire-diagram device editor (no FX)
Compressor & EQ presets	Yesbuilt-in + save your own	VoiceMeeter has macros	No	No	No
Per-source pan	Yesper incoming sender, in the hub mixer	YesVoiceMeeter	Via Dante Controller	No	Yes
Mismatched sample rates	Auto-resampledwarns when conversion is active	Manual rate match	Native	Native	Yes (CoreAudio handles it)
Remote mute	Yesper receiver and per source, from the hub	VoiceMeeter, local only	Via Dante Controller	No	Local only
Web dashboard for routing	Built-in hub	No	Dante Controllerdesktop app, paid	No	Native macOS app, local only
Browser-playable stream	AAC over HTTP per sender	No	No	No	No
Latency	~20 ms typical	Low single-digit ms	Sub-ms, broadcast-grade	Higher (video-first)	Sub-ms (no network)
Audio quality	Uncompressed 16-bit PCMup to 48 kHz, stereo	Uncompressed PCM	Up to 32-bit / 192 kHz	Uncompressed	Bit-perfect (CoreAudio passthrough)
Mixed sample rates in one mix	Yeseach sender can run any rate; receiver auto-resamples on the fly	No, single rate per session	Possible via clock domains	Possible, but heavier	Yes (CoreAudio resamples)
Open source	No (commercial)	No (free, closed)	No (proprietary)	SDK is free, not open	No (commercial)
Best for	Streaming, podcasting, project studios, live monitoring, multi-room playback, classrooms, dev/test	Streamers, all-Windows setups	Broadcast, large installs	Video production workflows	Mac-only routing, virtual cables, app capture

Tried to keep this honest — if anything's wrong or out of date, tell us. AudioBridger isn't the right tool for every job and this table should help you spot when it isn't.