What it is

Engage Ops puts live screens, windows, and cameras onto a display wall over the network. Someone at a desk shares their screen; a second later it is on the wall. An operator decides what goes where. Nobody runs a cable, and nobody installs a capture card.

It is built for rooms that stay on: command centers, network and security operations, dispatch, briefing rooms, live event production.

There are three jobs to be done with it, and this manual is organized around them. You probably only need one.

The pieces

One computer drives the wall. This manual calls it the Engage Ops host. It runs everything: the web console people sign in to, the signaling that introduces contributors to the wall, and the renderer that actually paints the displays.

Operators work in a browser, or in the Control App — the same console in an installed window, without the browser's certificate warning.

Everything else is a browser or an app on somebody's machine — with one exception. A machine with a camera or capture card can run the Capture Device Server, which offers its inputs to the wall for an operator to choose from, with nobody sitting at it. The host runs one for its own hardware automatically. See Capture Devices.

Canvas, tiles, sources, widgets

Four words appear constantly. They are worth two minutes.

Source

One thing somebody is sending: a screen, a window, a region of a screen, or a camera. One person can send several.

Canvas

The operator's map of the wall. What you arrange in the console is a scaled-down picture of the real displays, so a tile in the top-left of the canvas appears in the top-left of the wall.

Tile

A source placed on the canvas. A source that is connected but not yet placed has no tile and is not on the wall.

Widget

Something Engage Ops generates itself rather than receiving from a contributor. Today that means clocks. Widgets are always available and cost you nothing — see below.

The distinction that trips people up: connected is not the same as on the wall. A contributor can be successfully sharing and still see nothing on the wall, because the operator has not placed them yet.

How much fits

Nine sources on the wall at once, plus four widgets. That is the number to plan a room around.

Separately, twelve sources can be connected. The gap is deliberate and useful: three contributors can be signed in and streaming, ready to go up the moment the operator wants them, without occupying space on the wall.

First, connect to the console

Everything an operator does happens in the console, so start by getting into it and confirming the wall is running.

1. Open the console

   On the host machine, or any machine on the same network, browse to:

   https://<host>:8443/control/

   From the host machine itself you are signed in automatically. From anywhere else you will be asked for the Control UI password, which on a new system is password.

   Change it as soon as you have finished looking around — see Access control.

   Seeing a certificate warning is expected on a fresh system. You can click through it for now — see Getting past the certificate warning below for the two clicks. Certificates in Part 04 explains how to make it stop happening at all.

2. Check the wall is alive

   Look at the physical displays. An idle Engage Ops wall shows a welcome screen with the logo and the address contributors should visit. If the displays are black or the welcome screen only covers one monitor, set up display placement before going further.

3. Find the PIN and the share link

   In the console, open Settings → Control → Source sharing. You will see a five-digit Source-client PIN and a Source page URL, with a Copy button.

   https://<host>:8443/source/

   Send both to whoever is sharing. That is the whole invitation.

Your first source

With the console open and the PIN in hand, put something on the wall.

1. Share a screen

   On the contributor's machine, open the source page, enter the PIN, and choose Add Source. Pick a screen or a window and confirm. Their tile now says Streaming.

2. Put it on the wall

   Back in the console, the source appears in the source list down the side. Drag it onto the canvas. It is on the wall before you let go of the mouse.

   If it went up by itself, Auto-place new sources is switched on under Settings → Control → Streams.

Getting past the certificate warning

On a system whose certificate has not been set up yet, Edge and Chrome both block the page with Your connection isn't private and the code NET::ERR_CERT_AUTHORITY_INVALID. Your browser is saying it does not recognize who issued the wall's certificate — which, on a brand-new install, is true and expected.

It takes two clicks to continue. They are the same in Edge and Chrome; only one word of the link text differs.

Click here

Click here

Browser or app

There are two ways to share. The browser needs no install and handles most situations. The Windows app does more.

Use the browser if you are sharing once, from a machine you do not control, or from a Mac or Linux box. Use the app if you share regularly, need system audio, or need to send a region rather than a whole screen.

From a phone or tablet the browser is the only option, and it sends a camera or a microphone rather than a screen — see From a phone or tablet.

Getting in

Three ways, and the operator decides which are available.

With the PIN

Enter the five-digit PIN and you are in immediately. This is the normal path. If you do not have it, the operator reads it off the top of the Settings pane in Engage Ops Control.

With no sign-in at all

From the wall machine itself, or from an address the operator has put on the trusted list, there is no PIN and no waiting: give a name and you are on. Somebody sharing from the machine the wall runs on is already standing at the wall, so reading a PIN off the screen in front of them is a pointless step.

Request to join

No PIN? Choose Request to join, give your name, and wait. The operator gets a notification and approves or declines. You stay on a waiting screen until they answer; approval connects you automatically.

Requests expire after about five minutes. If yours lapses, submit again.

From a browser

1. Open the link and sign in

   Go to the source page URL and enter the PIN, or request to join.

2. Add a source

   Choose Add Source. Your browser asks what to share — a whole screen, an application window, or a browser tab. Pick one and confirm.

   To send a camera instead, pick the camera under Video input. To send only sound, choose Audio only.

3. Check it says Streaming

   Your card should show Streaming. If it says Waiting, the operator has not placed you on the wall yet — you are connected and nothing is wrong.

4. Add more, or stop

   Repeat for a second screen or a camera. Disconnect all takes everything down at once.

From a phone or tablet

The same page, and the same PIN. What changes is what you can send: phone and tablet browsers have no screen-sharing capability at all — the operating system does not offer one — so the page offers what the device can do, a camera or a microphone, and says in one line why the rest is missing. Nothing is broken and there is no setting to find.

Cameras are listed by which way they face rather than by name, with the back camera first: holding a phone up to show the room, a whiteboard or a piece of equipment to the wall is the common case, not a selfie.

From the Windows app

Connecting the first time

Launch the app and it asks you to Connect to your Engage Ops system. Enter the address of the wall, then the PIN. Once accepted the app shows Paired and remembers the system — next launch it reconnects on its own.

Picking what to send

Add source opens the picker, which has three sections:

Displays

Your monitors, with live thumbnails.

Application

Individual open windows, searchable by name.

Devices

Cameras and capture cards, with a live preview. Each one lists every format it advertises — pick the resolution and frame rate you want, and which audio input goes with it.

Click a tile or Full to send the whole thing. Choose Area to drag out a region — only that rectangle goes to the wall.

Sharing straight from the taskbar

On Windows 11, while the app is signed in, hovering any window's taskbar button shows its thumbnail with a Share this window button on it — the same place Teams puts one. Click it and that window goes to the wall without opening the app's picker at all: it is added with the picker's defaults, 30 fps with audio and the cursor included. The thumbnail then reads Stop sharing, which removes it again. A window that is already a source will not be added twice.

Reading the status badges

Each source card carries a badge. This is the most useful thing in the app, because it tells you whether you are actually on the wall.

Sending audio

The Windows app can send your system sound, a microphone, or both mixed together, as a single audio stream. You can add and remove inputs while you are live, and set gain or mute each one independently.

Audio is carried at full stereo quality rather than the voice-optimized, heavily processed audio a conferencing tool would send, so music and alarm tones come through as they actually sound.

A source that sends only sound still gets a tile if the operator places it. It reads Audio Only over a sound wave that moves with what the wall is playing from you — so the room can see at a glance that your feed is alive, and that it is the one being heard.

Quality settings

The defaults are correct for almost everyone. Engage Ops adapts resolution and frame rate automatically when the network tightens. Change these only for a specific reason.

Resolution

Caps how large a picture you send. Lower it if your upload is constrained.

Frame rate

Leave high for video and moving dashboards; lower it for a static page to free up bandwidth for detail.

Video codec

How the picture is compressed. Leave on the default unless you are troubleshooting with support.

Encoder

Which piece of hardware does the compression. Auto picks for you. Choosing a specific graphics card can help on a machine with more than one.

Privacy and pausing

Privacy blanks your source on the wall without disconnecting it. The tile goes black, your audio goes silent, and your slot stays reserved. Turning it off puts you back instantly — there is no reconnection and no gap.

Use it when you need to check email, enter a password, or handle something you would rather not put on a wall in a room full of people.

The operator has an equivalent control for the whole wall; see Blank Screen.

When your network drops

You do not have to do anything. Both clients reconnect on their own, waiting a little longer between each attempt, and your sources return to the same places on the wall.

This covers dropped Wi-Fi, an unplugged cable, a disabled adapter, a VPN reconnect, and the wall itself restarting.

What does not come back by itself:

• A browser tab that was closed — open the page again.
• A machine that went to sleep — wake it and check the client.
• A shared window whose application quit — the app marks it HELD; pick the window again.

The console

Open https://<host>:8443/control/ from any machine on the network — 8443 unless an administrator has moved the port. From the host itself you are signed in automatically; from anywhere else you need the Control UI password. On a system nobody has configured yet that password is password, which your administrator should have changed — if it still works, tell them.

Several people can have the console open at once. Everyone sees the same canvas, and a change one operator makes appears immediately for the others.

Open it on a phone and you get a different interface built for a thumb — see The phone remote.

You can move the source list to the left, right, or bottom edge under Settings → Control → Source list.

The phone remote

The same address on a phone gives you a wall remote, not the desktop console squeezed onto a small screen: the things you do standing in front of the wall, laid out for one hand. It talks to the same system as the desktop console, so anything you change on it appears on every other operator's screen at once.

Which interface a browser gets is decided by what you are pointing with, not by how wide the window is. A touch device with a small screen gets the remote; a desktop window dragged narrow next to another application is still driven by a mouse, so it keeps the full console and simply rearranges itself.

What is on it

Wall

A live picture of the wall with every placed source, and the list of what is on it. Tap a source to manage it. In Freeform you can also arrange it with your fingers: drag to move, pinch to resize. The gesture works anywhere on the preview, not just on the tile, because a tile that small has no room for two fingers. A locked tile can only be selected, not moved. The Freeform / Grid switch and the blank-screen button are here too.

Sources

Everything connected, placed or not, and the way to put one on the wall.

Environments

The same saved layouts as the desktop — which is the point of carrying this in your pocket: you can load Incident from the doorway.

More

Users, Settings, sign out, and Switch to full console for this device.

Tapping a source opens a sheet with its controls: add to the wall, spotlight, mute, name label, lock, permanent, remove. In Grid you also get a cell picker and span controls; in Freeform, snap presets, nudge arrows and smaller / larger.

What it deliberately leaves out

Staging, the projection and renderer-placement editors, monitor alignment, the logs timeline, and the grid's own size are setup and precision work. They stay on the full console, where there is room to do them properly.

The colors of your own console

Three more settings in Settings → Display → Control UI change how the canvas looks to you, and never touch the wall: Preview canvas color fills the wall preview, Letterbox color fills whatever your wall's shape leaves uncovered around it, and Guides color draws the dots and grid lines, following the canvas color by default so the guides stay faint against it. Type a hex value (#1a2b3c) or R, G, B numbers, or click the swatch and pick. Like the theme, these are remembered on that device only — a dark room and a bright office can each have what suits them.

Choosing the interface yourself

Settings → Display → Control UI → Interface: Automatic, Full console or Phone remote. Automatic is the rule above. The choice is remembered on that device only, so putting one phone on the full console does not change anybody else's. The remote's More tab has the same switch.

The Control App

The console in a browser means a certificate warning on every machine that has not been given the system's root certificate — see Getting past the certificate warning. The Engage Ops Control App is the way around that for the people who operate a wall every day: an installed Windows application that runs the same console in its own window, with no warning to click through and nothing to import into Windows.

It is the operator's counterpart to the Windows source app, and it is a control surface only — there is no + Add Source button in it. Sharing still happens from the source app or the source page, which is what the hint in its place says.

Signing in

1. Pick the system

   Find systems on my network lists the walls that answer on the local network; or type an address if discovery is not available to you. On the wall machine itself there is nothing to sign in to — it is already trusted, as the browser console is.

2. Enter the controller password

   The same password the browser console asks for. Remember password is on by default and keeps it in Windows Credential Manager, so later launches show Saved password and an empty submit signs you in.

3. After that, it opens on the wall

   A remembered system reconnects at launch and goes straight to the console. Sign out ends the session but keeps the saved password, the way a browser does; the × on a saved system forgets both.

Installing it

Copy its folder to the operator's PC — somewhere the signed-in user can read, such as %LOCALAPPDATA%\Programs\EngageOps\Control\, which needs no administrator rights — and make a shortcut to engage-ops-native-control.exe. The ui folder beside it must stay there. There is no installer, and no runtime to add: Windows 11 already has the WebView2 component it needs, and on Windows 10 you may have to install Microsoft's Evergreen WebView2 runtime if the window opens blank.

It needs the control port (8443 by default) to the system. Find systems on my network also needs UDP 47820 on the local network; typing the address works without it.

Freeform and Grid

The wall is always in one of two modes, chosen at the left of the toolbar.

In Grid mode, Cols and Rows fields appear in the toolbar. Change them and the wall re-flows immediately.

Placing sources

Drag a source from the list onto the canvas. In Freeform, drop it where you want it and drag its edges to resize. In Grid, it snaps to a cell.

To take something off the wall, use Remove on the tile. The contributor stays connected — they return to the source list and can be placed again instantly.

Auto-place

Settings → Control → Sources → Auto-place new sources puts every new arrival onto the canvas automatically. Good for an open collaboration session; switch it off when you want to vet what reaches the wall.

In Freeform it does more than place the newcomer. While it is on, Engage owns the arrangement: every source joining or leaving repacks the whole wall into a space-filling layout. You can still drag and resize by hand, but the next repack overwrites that — use Lock on a tile to hold its position and size while the rest keep repacking. In Grid, a newcomer takes the first free cell and the tiles already placed stay where they are.

Sources that disconnect

When a contributor drops, their tile does not vanish. It becomes a placeholder reading Waiting for connection, holding the spot, so that when they reconnect they land back exactly where they were.

To clear those out automatically, turn on Expire missing sources and set a timeout under Settings → Control → Sources. Off by default — a placeholder is held until you remove it. With it on, the placeholder shows a live countdown to its own removal.

Tile controls

Hover a tile on the canvas and its controls appear. On a small tile they collapse behind the … button.

When a source stops updating

A source that has stopped producing new pictures does not look any different on the wall: capture keeps re-sending the last frame it had, so the tile is a perfectly healthy stream of a picture that may be minutes old. Nothing at the wall can work that out for itself, so sources report it, and the console puts a badge on the card and on the tile.

The badges come in two kinds, and the difference is deliberate.

Something is wrong, and we know what

Not updating

The captured window has been minimized. It produces no frames while it is, and the wall is showing the last picture it made. Ask the contributor to restore the window.

Not responding

Windows reports the captured application as hung. The wall is showing the last picture it made before it stopped answering.

No desktop

The source machine has no desktop to capture — a remote session was disconnected without signing out. Nothing composites there, so no capture of any kind can produce a frame. Somebody has to sign back in to that machine.

Region held

The display a region was drawn against is no longer at that resolution, so the source is holding the region rather than streaming whatever reflowed into it. The contributor reselects the region.

A statement of fact, not a fault

No new frames · 12 min

No new pixels for that long, and no reason available. This is exactly what a genuinely static source looks like — a dashboard, a slide, a status board — so it is reported without alarm. If that source should be moving, it has stopped; if it is a wall-mounted duty roster, it is doing its job.

Private

The contributor has privacy on and the wall is showing a PRIVATE card. Real pixels never leave their machine while it is on.

Times are coarse on purpose — "a few seconds", "5 min", "1 hr 20 min". They are read to answer is this stuck, and they are stamped by the source's own clock, not the wall's.

The source list

Everything connected, split into Sources and Widgets, each with a count so you can see how close to the limit you are.

• + Add Source opens the source-join page in a new tab — a quick way to put something up from the operator machine itself.
• + Add Capture Device puts a camera or capture card that is plugged into some other machine onto the wall. Once you have some it reads Manage Capture Devices. See Capture Devices.
• Search matches on source name, contributor name, or internal ID.
• Sort reorders the list; useful once you are past a handful of sources.

Two bulk buttons sit at the top of the panel. They are not the same button with different wording, and the difference is the whole point:

Remove All

Clears the layout you are looking at. Tiles come off the canvas; every source stays connected and can be placed again from the panel whenever you want. The other layout is untouched, and pinned (Permanent) tiles are left where they are — which is why the button grays out on a layout holding nothing but pinned tiles.

Delete All

Disconnects every source and clears both layouts. It takes pinned sources too, and says so before it does. Widgets are left alone. You are asked to confirm, and the confirmation holds the button for a moment on purpose.

Capture Devices

A Capture Device is a camera, capture card or audio input plugged into some other machine, put on the wall by you rather than by whoever sits at that machine. Nobody has to be at the far end: the operator picks the device and it starts streaming.

A machine offers its inputs by running the Capture Device Server, and there are two ways that happens:

This system

The Engage Ops host offers its own hardware with nothing to install and nobody to log in — it is started and supervised along with everything else. A capture card in the wall machine is available the moment the system is up.

Any other PC

Install the Capture Device Server on it. Somebody signs it in to this system once, the same way a contributor signs in, and from then on its inputs are offered to you. It has a small window listing what it is currently sending and a Close button; closing it does not take anything off the wall.

+ Add Capture Device in the source panel opens the picker. It lists every machine that is offering devices, grouped by machine, with the connection slots you have left across the top.

Online / offline

Each machine says which it is. An offline machine stays listed rather than vanishing — its sources are still on your wall — but nothing under it can be added while it is away. Remove this machine is there for one that is never coming back; it forgets the machine and deletes its sources.

Checking which devices are free

The machine is asked what it can actually open. A device another program already holds comes back unavailable, with the reason the machine gave.

Capture mode and audio

Pick the format the device should run at, and whether to bring in the audio paired with it, a different input, or none. Both can be changed later from the device's card.

Add one and it becomes a card in the picker's first pane and a source in the panel — connected, but not placed. Drag it onto the canvas like anything else. Rename it, change its mode, or take it off the wall from its card.

Approving requests

If Request-to-Join is enabled, contributors without the PIN can ask to share. A chip appears in the toolbar when someone is waiting; click it to see the requests, each with the name they gave.

Approve and they connect automatically. Decline and they are told. Requests expire on their own after about five minutes.

To turn the feature off entirely and require the PIN for everyone, use Settings → Control → Request-to-Join. The same place holds rate limits that stop a single machine flooding you with requests.

Blank Screen

Blank Screen in the toolbar hides every source on the wall at once. The room sees nothing; you keep your previews in the console and can carry on arranging. Click again to bring it all back.

Use it when someone walks into the room who should not see what is up, or while you rebuild a layout you would rather put together unobserved.

Designing what a blanked wall shows

By default a blanked wall falls back to its ordinary background. It can show something of your own instead — a holding card, a logo, Session in progress — which is better than a blank wall in a room people walk into and wonder whether something is broken.

Switch on Settings → Display → Blank screen → Custom blank screen, then open Customize. You build the screen out of layers on a preview of your wall:

A background

One, underneath everything, and the same choices the wall background offers: a gradient, an image, or an animated shader.

Up to nine text and image layers

In any mix. Text has its own size, color and alignment; an image has a choice of how it fits its box, and a PNG keeps its transparency, so a logo sits directly on whatever is under it. Both can be rotated and given a shadow.

Placing them

Drag a layer on the preview, or nudge it a wall pixel at a time with the arrows. Layers snap to each other's edges and corners as you drag, with guides showing what they lined up against — including layers you have rotated to the same angle.

Nothing reaches the wall until you choose Apply to wall, so you can drag a caption around without the room watching you do it. If the wall is blanked while you are working, Apply shows the result immediately.

Widgets

Widgets are sources Engage Ops generates itself. Today that means clocks. They behave like any other source on the canvas — drag, resize, place — but they are always available and cost you neither one of the nine on-wall slots nor one of the twelve connection slots.

Widgets have a pane of their own. Two ways into it:

Settings → Widget

Reads back how many widgets are defined and how many are switched on. The button opens the pane.

The gear on a widget's card in the source list

Opens the pane on that widget, already selected — you have already said which one you mean by clicking it.

The pane lists every widget down the left, each with a dot showing whether it is on, and + Add date & time widget below the list with a count beside it. Four is the limit. Select one and the right-hand side is that widget in full:

Name

What the source list calls it, and what the on-wall name plate shows.

On / Off

Off stops generating it: it leaves the wall and the source list. The definition stays, and stays editable — setting a widget up before switching it on is the normal order of things.

Delete

Removes the definition, and the widget with it. You are asked to confirm. To keep a widget but take it off the wall, switch it off instead.

Time zone and caption

The point of having four: a row of clocks for the regions you cover, each captioned with its city.

Clock

Digital or analog, with optional seconds and date, and a choice of date format.

Appearance

Tile shape, font, bold, and colors for the time, for the date and caption, and for the background.

Under the name, one line says where that widget currently is — switched off, starting up, waiting in the source list, or placed on the wall — so a widget you have not placed yet is never mistaken for one that is not running.

Staging

Staging lets you build a layout without the room watching. Click Staging, choose whether you are staging a Freeform or a Grid layout, and arrange it off-air. Commit it and it replaces the live layout of that type and becomes the active mode.

This is the tool for a planned transition: build the next configuration during the current briefing, then switch in one action.

Environments

An environment is a saved layout you can bring back later: the mode, every placement, the grid size, and the per-tile settings such as pinned, labeled, and which source owned the audio. Thumbnails are saved with it, so the list is browsable months later.

Save the layouts your room actually uses — Morning brief, Incident, Handover — and load one instead of rebuilding it under pressure. The pane closes itself once a save lands, putting you back on the wall you just captured; if the name is already taken it stays open and asks.

What an empty wall shows

Four things can occupy a wall with nothing placed on it, all under Settings → Display.

Welcome screen

The default. Shows the logo, the address contributors should visit, and a QR code for it, so anyone walking into the room can get onto the wall without typing anything. Customizable — see below.

Wall background

What sits behind your tiles. Set it to something the room reads well against.

Screen saver

Kicks in after a set number of idle minutes. Several styles, including a plain clock on black.

Canvas test pattern

For commissioning. Turn it on to check every display is aligned, correctly ordered, and showing the right part of the canvas — then turn it off.

The QR code

Under Settings → Display → Welcome screen → Customize, alongside the wording and the logo. Someone in the room points a phone at it and lands on the source page; on a phone that means a camera straight to the wall, with nothing to install and no address to read across a room.

In the card

The default: under the address, in the middle of the screen. Best on a single display.

Custom position

Drag it in the preview to wherever works best in the room, or select it and use the arrow keys. The card shortens and recenters to stay out of its way.

Move around the screen

Every ten seconds the code disappears and comes back somewhere else in the empty space around the card. Useful on a big wall where one fixed corner is awkward to reach. It overrides the position above.

Live View

Live View streams the whole composited wall to a browser, so people who are not in the room can watch what the room sees.

It is one view of the entire wall, not individual sources, and it is view-only — a Live View viewer cannot rearrange anything.

Requirements

Installing

Engage Ops is delivered as a Windows installer. Run it on the machine that will drive the wall and accept the defaults — there is nothing to decide during setup. Everything that needs configuring is configured afterward, from the console.

When it finishes, go straight to the first-boot checklist. That is where the real setup happens.

First-boot checklist

In order. Each step assumes the one before it.

1. Open the console on the host

   Browse to https://localhost:8443/control/ on the host machine. Connections from the host are trusted automatically, so you are signed in without a password — which is how you get in before a password exists.

2. Change the Control UI password

   A new system ships with the password password. Anyone who can reach the host on the network can sign in and take over the wall with it.

   Change it at Settings → Control → Control UI password. Leave Current password as password, set the new one, and confirm.

   Do this before the host goes onto a network anyone else can reach. While the default is still in effect the console keeps reminding you, including on the host itself.

3. Place the renderer across your displays

   Settings → Display → Renderer window placement. This is the step that turns a pile of monitors into a wall. See Display placement.

4. Set up the certificate

   Either import a certificate from your own authority or generate the default one, then distribute the root to client machines. See Certificates. Doing this now saves every contributor a scary browser warning later.

5. Set the public address

   Settings → Network → Public address → Public URL. This is the address Engage Ops hands out in share links and shows on the welcome screen. Without it, the console guesses from your own browser's address bar, which may not be reachable from where your contributors are.

6. Note the PIN and test a share

   Settings → Control → Source sharing. Share a screen from a second machine and confirm it reaches the wall before you call the install done.

7. Decide the join policy

   PIN only, or PIN plus Request-to-Join. See Access control.

8. Only if contributors will come from outside this network

   Set up STUN and the TURN relay, and prove the relay works before you need it — a relay that was never tested is indistinguishable from one that was never configured. Reaching the wall from outside. Skip this entirely if everyone shares from the same network as the wall.

Display placement

The renderer runs as one borderless window stretched across the displays that make up the wall. You define that window in Settings → Display → Renderer window placement, using a visual editor showing your monitors as the system reports them.

The placement is remembered along with a record of the monitor arrangement it was set for. If the arrangement later changes — a display is unplugged, reordered, or its resolution altered — the saved placement no longer matches and the renderer falls back to covering the primary display only.

Wall audio output

Whichever source currently owns the room's audio, the wall plays it through one output device, configured under Settings → Audio.

Device

default lets Windows pick, which is almost always what you want. Name a device only to pin the wall to a specific output — an audio interface that must be used regardless of what Windows has decided is default this week.

Sample rate

48 kHz is what the media transport carries, so nothing is resampled. Higher rates buy no quality for streamed audio and cost CPU.

Channels

Two for ordinary stereo playback.

Jitter buffer (ms)

How much audio is held before playing, to absorb network jitter. It is also how far audio lags the wall's video, because the render loop does not compensate for it. 40 ms is the tuned default — raise it only if audio breaks up, and expect visible lip-sync drift well before you reach the maximum.

Certificates

This is the single most common source of support calls, and it is worth twenty minutes on install day.

Everything runs over HTTPS. A browser trusts an HTTPS server only if the server's certificate was issued by an authority the browser already trusts. A fresh Engage Ops host issues its own, from its own private authority, which no machine has heard of — so every browser warns.

Everything below is in Settings → Certificate.

Option A — use your own certificate

If your organization issues certificates, use one. Under Import a certificate, supply the certificate file, the key file, and the key password if it has one. Applying it restarts the control service: your console page disconnects for a few seconds and needs a reload. The wall keeps displaying throughout — the renderer is never restarted for a certificate change.

Make sure the certificate covers the name people will actually type. A certificate valid for the wrong name warns in exactly the same way as no certificate at all. The Valid addresses section lists the URLs your current certificate genuinely covers, and flags the ones that will still warn.

Option B — use the built-in authority

Generate default certificate creates a private authority and a server certificate from it. Then, under Certificate for client, use Download root or Copy download link and install that root on every machine that will view or publish to the wall.

On Windows: open the downloaded file → Install Certificate → Local Machine → Trusted Root Certification Authorities.

Network and ports

Public URL

Settings → Network → Public address. Set this to the address contributors should use. It is what share links contain and what the welcome screen displays. Leave it unset and Engage Ops falls back to guessing from the operator's own browser — often an address that only works inside one subnet.

Changing the port

8443 is a default, not a fixture. Settings → Network → Server holds the addresses this system answers on.

Control server port

The port this console, the source clients and every paired device reach the system on. The button reads Apply and restart: Engage Ops restarts, and the console moves itself to the new address — a banner at the top of the window carries the new link. The console will not accept a port browsers block outright, warns on ports below 1024, and refuses to let the control and signaling ports collide. If a Public URL is set, it wins over this port in every address Engage Ops hands out, Live View included — so change its port to match, unless a router or reverse proxy maps one port to the other. The console warns beside this field while a public URL is set.

Redirect http:// to https://

Offered only when the control port is 443. Answers plain http:// on port 80 with a redirect to this system's https address, so someone who types the bare host name lands on the portal instead of a browser error. Port 80 has to be free — Engage Ops checks when you apply and will not take it from another program. On any other port there is no https address an http:// caller could have guessed, so the option is not offered. A port and a protocol cannot share one socket: http://host:8443 can never be redirected to https://host:8443.

Signaling port

Internal, loopback only — browsers never reach it directly, they go through this system's own address. Change it only if something else on the machine already holds the port.

Diag interval (ms)

How often CPU, memory, disk and network figures are sampled for the diagnostics overlay. Applies immediately, no restart.

When peer-to-peer is blocked

Media normally flows straight between contributor and host, and on one network that always works. Contributors coming from outside it may need help, and sometimes a relay. That takes some setting up — see Reaching the wall from outside.

Reaching the wall from outside

Audio and video do not travel through the console or the web server. Each contributor's media goes straight to the wall machine, which is why one host can carry a dozen sources. For that to happen the two ends have to find a path to each other, and on a single network they always do.

From outside — another site, a home connection, a guest VLAN, a phone on cellular — both ends sit behind routers that hide their real addresses. Engage Ops has two mechanisms for this, and they are not alternatives: STUN finds a direct path, and the TURN relay carries the media when there is no direct path to find. Both are under Settings → Network.

STUN — finding a direct path

A STUN server does one small thing: it tells whoever asks what their address looks like from the outside. Both ends ask, trade the answers, and try to meet in the middle. It is cheap, it is quick, and when it works the media goes direct, with no load on your host at all.

Settings → Network → STUN holds the list that every client, and the wall itself, is given:

The built-in servers

Two Hiperwall-operated servers, marked built in. They cannot be removed, but the checkbox beside each one takes it out of use — prefer that to deleting, so you can fall back later.

Adding your own

stun:host:port, or just host:port; the port defaults to 3478. Host names, IPv4 addresses and bracketed IPv6 addresses are accepted, and anything else is refused with the reason. Up to eight of your own.

What several of them do

They are asked all at once, not in turn. An extra server is a backup rather than a delay, and one that is down costs nothing. There is no reason to prune the list for speed.

Changes apply to new connections immediately — nothing restarts.

The TURN relay — when there is no direct path

Some networks refuse direct connections between peers no matter what STUN discovers. A relay is the fallback: media goes to the relay, and the relay passes it on. Engage Ops ships one, off by default, under Settings → Network → TURN relay.

Clients reach for it only after a direct path has failed, so turning it on does not push your local contributors through it. What it costs is the host's bandwidth and a little latency, for the contributors who need it.

Setting it up

In this order. The first two are what people most often leave out, and both fail quietly.

1. Turn on Enable relay

   The Status block below should show a port count within a few seconds. If it stays on Not running, the Logs pane has the reason under engage-turn.

2. Set the relay public address

   Relay public address, under Advanced. If the host sits behind NAT with port forwarding — the usual reason to be doing any of this — this has to be the public address clients can actually reach. Leave it blank and they are handed the machine's internal address, which from outside points nowhere. The console flags it while it is empty.

   A system with neither this nor a Public URL set cannot name itself at all, and the relay is then quietly left out of what clients are given. The log says so; nothing on screen will.

3. Open the firewall

   Inbound to the wall machine: 3478 UDP and TCP for the listener, 5349 TCP for the secured listener if you keep it, and the relay range — 8081–8104 UDP by default. Changing the numbers in Settings does not touch your firewall, and a relay whose ports are closed fails silently: the contributor simply never connects.

4. Check the port numbers suit the machine

   The listener and the secured listener must stay outside the relay range or they collide, and the console refuses to apply that. Keep clear of whatever else runs on the machine — the range starts at 8081 rather than 8080 because 8080 is this system's own signaling port.

5. Apply & restart

   This restarts the relay, not the system. Anything currently relaying reconnects; the wall itself is unaffected.

6. Prove it works

   On your own network nothing ever relays, so a misconfigured relay looks exactly like a working one. Turn on Force relay for all clients under Diagnostics, connect a source, and confirm it comes up. Then turn it off — while it is on, every local contributor is pushed through the relay too.

Sizing the relay range

A relaying source holds one relay port for as long as it is connected, so the twelve sources this system supports need twelve ports at once. The default range is twenty-four: twelve, and the same again as headroom. Set fewer than twelve and the console says so in red but still lets you apply it — it is your call, and the cost is sources that fail to connect under load.

The headroom is there for ungraceful disconnects. A client that closes cleanly gives its port back at once; one that vanishes — a dropped link, a killed browser tab — leaves its port held for up to ten minutes. That ten minutes is fixed by the relay itself and cannot be shortened, so a contributor on a flapping connection can be holding several ports at once. If you run short, widen the range; there is no timeout to tune.

Relay ports in use in the Status block is the number to watch. It turns red as it approaches capacity, and the console says so outright once allocations have been refused for want of a port.

The secured listener

The secured listener carries relay traffic over TLS on TCP, which is the only thing that gets through a network permitting nothing but outbound TLS — exactly the networks whose users need a relay in the first place. It uses this system's own HTTPS certificate, so the client has to trust that certificate already: on a system still using the built-in authority, that means the root certificate has been installed on the client machine. See Certificates. Set the port to 0 to switch the secured listener off.

Reading the status block

Relay ports in use

How much of the range is held right now, against its capacity.

Allocations refused

Requests the relay could not satisfy. It means the range is too small for the load — widen it.

Refused by the peer allow-list

Requests to relay traffic somewhere this system does not permit. Expected to be zero on a normal system; anything else is worth asking who is pointed at your relay.

Credential rejections

Clients presenting credentials the relay would not take. A handful around a restart is ordinary — the credentials handed out are short-lived by design.

Discovery

Discovery lets clients find walls on the network rather than being handed an address. Configure it under Settings → Network → Discovery.

Discoverable

Whether this host answers discovery queries at all.

System name

How this wall identifies itself. Name it after the room.

Site

Groups walls in the same building or campus.

Visibility

Who may discover it.

Site key

A shared secret restricting discovery to clients that know it.

Discovery status shows whether the host is currently responding, and on which UDP port. Search this network below it asks the network the same way a client does and lists what answers: use it to confirm this host is reachable, to catch two walls sharing a name, and to catch a copied installation answering with the same system ID under a different certificate.

Access control

Who can operate

Connections from the host itself are trusted automatically and never prompt for a password — that is how you get in on install day. Everyone else needs the Control UI password, set under Settings → Control → Control UI password. The Users button in the toolbar manages operator accounts, and holds the banned and trusted address lists that used to sit under Settings.

Who can share

The five-digit PIN, under Settings → Control → Source sharing. It is also shown in the header of the Settings pane whatever tab you are on, so you can read it out to somebody without navigating anywhere. Regenerate it whenever you want to cut off future joiners; already-connected contributors keep working, so regenerating mid-session does not knock anybody off the wall.

Request-to-Join, on the same tab, lets people without the PIN ask an operator for access. Turn it off to require the PIN for everyone.

Contributors on the wall machine itself, or on a trusted address, skip both: they give a name and connect. Trusted addresses are a deliberate grant — see the table below for where they live.

Hardening

Files and locations

Two trees. Configuration and state live in one; logs in the other.

Restart and power

Settings → Control → Control holds four actions for the host, so you do not need a remote desktop session for routine recovery:

Restart Engage Ops System

Restarts the software only. The machine stays up. The console reconnects by itself. This is the first thing to try.

Restart machine

Reboots the host. Sources reconnect and return to their places afterward.

Turn off display

Puts the wall's panels to sleep. For overnight, or an empty room.

Wake display

Brings them back.

Symptoms and fixes

Nothing on the wall

A source will not join

A source is connected but wrong

Everything stopped at once

Logs

Logs in the toolbar opens the log viewer and timeline without needing access to the host's filesystem. Start here.

On disk, logs live under a separate tree from configuration, one folder per component:

The console can export a log bundle as a single zip — that is the right thing to attach to a support request, because it captures every component's view of the same moment rather than just the one you suspected.

The Diag toggle overlays live performance figures on the canvas — useful for confirming whether a stuttering source is a decode problem at the wall or a network problem at the contributor.

It also lists the host's graphics hardware. Collapsed, the row names the GPU rendering the canvas; expanded, it shows every adapter with its 3D and copy engine load and its temperature, and marks the one driving the wall.

What recovers itself

Engage Ops is designed for rooms that cannot go dark, and a great deal of recovery is automatic. Knowing which half is which saves you intervening in something that was about to fix itself.

Limits and specifications

Glossary

Canvas

The operator's scaled map of the wall.

Control App

The console as an installed Windows application, with no browser certificate warning.

Capture Device

A camera, capture card or audio input on another machine, put on the wall by the operator rather than by anyone at that machine.

Capture Device Server

The small agent a machine runs to offer its inputs to an Engage Ops system.

Contributor

Anyone sharing a source.

Engage Ops host

The computer driving the wall.

Environment

A saved, named layout that can be reloaded.

Freeform

Layout mode where tiles go anywhere at any size.

Grid

Layout mode where tiles snap to an even grid.

Live View

Watching the whole wall from a browser elsewhere.

Permanent

A tile pinned so it cannot be removed by accident.

PIN

The five-digit code contributors use to join.

Privacy

A contributor blanking their own source without disconnecting.

Public URL

The address Engage Ops publishes in share links.

Request to Join

Asking an operator for access instead of using the PIN.

Source

One screen, window, region, or camera being sent.

Spotlight

Temporarily enlarging one source to fill the wall, keeping its shape.

Staging

Building a layout off-air, then committing it.

Tile

A source placed on the canvas.

Widget

A source Engage Ops generates itself, such as a clock.