01Check
Clear the rig first
System Check surfaces permissions, slide apps, displays, speech, and optional NDI before the room is waiting.
The model
Native decks in. One show state out.
Downstage does not convert your slides. It coordinates the operator surfaces around native playback.
Source
Keynote · PowerPoint · PDF
Original files, builds, video, and fonts
Downstage
Queue · launch · track
One shared show state
The room
Audience · talent · director
The right view on the right screen
Audience outputPresenter viewTeleprompterPhone directorClocks & cues
The workflow
From setup to GO in four stages.
Prepare the rig once, rehearse the real path, then stay on the queue.
02Build
Make the show queue
Add mixed deck types, set each output, and attach notes or scripts. Tracking preparation begins at load.
03Shape
Fit the room
Drag and resize presenter tiles, save the layout, then pair the phone director by QR code.
04Rehearse & run
Test the path. Call GO.
Run the actual displays and network, save the show for backup, then launch every cue from one surface.
The features
Six surfaces. One show.
Each solves a specific job in the room.
Show Queue
Order mixed decks, assign outputs, attach scripts, and launch each cue.
Free-form view
Move and resize slide, notes, teleprompter, and clock tiles for the room.
Speech-following text
Follow the speaker offline, with manual control and Recall when the route changes.
Light-table calling
Follow the live slide, zoom the grid, navigate behind ARM, and message the presenter.
Clocks & messages
Put timing and operator messages on downstage screens, not in the deck.
Show integration
Connect NDI, network cues, and Bitfocus Companion when the rig needs them.
Semantic Recall
Speech tracking that can find its place again.
Bundled Whisper follows the words on the page. Downstage prepares a location map at load; optional local Ollama models add stronger recovery after paraphrases, skips, backups, and asides.
If two distant locations are genuinely ambiguous, Recall stays put instead of guessing.
Offline speech tracking
→
Prepared local recall
→
Hold position
Founding beta
A small beta for real operators.
We are starting with 25 operators in waves of 8–10.
Good fit now
You can rehearse a real show.
- Apple Silicon Mac with presentation-heavy work
- Access to the intended displays and network
- Willingness to report what happens in the room
Better later
You need an appliance.
- Windows or Intel Mac support
- Zero rehearsal before first use
- Control across isolated venue networks
macOS, Keynote, and PowerPoint updates can change permissions or launch behavior.
Phone control needs a tested local network without client isolation.
Core tracking is bundled; strongest paraphrase recovery uses optional local models.
Founding beta · limited cohort
Tell us what kind of show you run.
Include your Mac, slide apps, display count, and next rehearsal.