Changelog

• The privacy curtain. Press K and the room goes blank — not held on the last frame, but cleared. Step away mid-talk and nothing from your screen is left on theirs. Press it again and the room comes back to where you are.
• Marked text stays home. Anything you mark data-pt-private in a captured page is masked before it ever leaves the tab, exactly as our disclosure promises.
• See it on a phone first. The Join panel now offers a phone-width preview of the deck, before you start.
• A warmer close. When the session ends, the presenter gets a recap that reads like the end of a talk — the reactions tile shows what the room actually sent — and the session keeps its name when you come back to it.
• Search, and plainer words. A search field in the console, an honest offline line for the audience, and data copy that says what we do in plain language.
• Faster to open. The loader gets the room in about 20% sooner.
• Fixed: an auto-animate transition no longer flashes a scrollbar, and a deck chosen after the session starts now follows on every screen straight away.

• One toast, and it clears itself. Notifications no longer stack up over your talk — a message appears, reads, and goes.
• Plain words. Errors and copy talk to you, not about the machinery: no protocol names, no status codes.
• Quieter loading. Skeletons instead of spinners, one empty state per screen, and plainer labels across the console and the admin.
• Your end screen. When a session ends, the presenter gets a recap of their own talk — not an operator’s readout.
• No accidental self-capture. The cockpit no longer offers the console’s own tabs as a source to capture.

• A projector view. From the cockpit, open a window on the venue’s screen. It shows the deck, letterboxed, with none of your notes, tools or counts — and it survives the things venues do: unplug the display mid-talk and it steps aside and comes back, the screen stays awake, and the cursor hides while you talk. If the room’s screen is driven by another machine, that machine can open the join link as a display instead.
• The cockpit is the session. Creating a session drops you straight into it. Set up your deck or a tab capture, then start from the Join panel and share the link — there is no separate start screen to get past.
• A capture that can’t take the room down. The capture now bounds itself, so a heavy page degrades gracefully instead of crashing the tab mid-sentence.
• Phones behave. The rail is a proper bottom sheet, and the page no longer zooms in on every input on iOS.

The cockpit has been rebuilt around one idea: the stage keeps the screen, and everything else defers to it.

• Rename and reshape a deck. A deck’s title and its canvas — the frame every pointer position is measured on — are editable from the deck’s card, without re-importing it and losing the slides.
• A slide move is now instant. The deck travels to the room once, as a file addressed by its own contents, and a slide change sends a hash and a position instead of the whole deck. A room of a hundred watching sixty slides used to move about 120 MB of repeated markup; it now moves a few kilobytes. A reconnect, a late join or a replayed recording fetches nothing at all.
• The room’s reactions float up your screen, scattered, one per reaction — the same drift an audience member sees on their own.
• A point-back is answered where it was made. When someone points at an element, the accept and decline marks sit on that element in your deck, as well as in the rail.
• Your deck’s navigation is below your deck, with the position between the arrows, and the tools are one centred pill.
• The cockpit fits a phone and a tablet. Below a tablet’s width the rail becomes a panel over the stage — a bottom sheet on a phone — and starts folded, so the stage is what opens. The view switch moved into the top bar, where it only appears when there is actually another view to switch to.
• Watching is quiet. In the audience view there are no controls, because none of them do anything there — just a small pill saying what you are looking at.

You can now present a deck, not only a captured tab. Bring a reveal.js markdown file or a reveal.js HTML export (a Slides.com export works too) and Present Tense imports it once, then renders it on every screen as real text.

• Import once, render everywhere. The deck is normalised at import — reveal’s own plugin does the conversion — so what is stored is plain reveal markup. What crosses the network is the deck itself, a few kilobytes, not a picture of it.
• The room follows you. Change slide and the room changes with you; point, draw or highlight and it lands on the same element on every screen.
• The room can point back. A viewer picks an element and you see it marked on yours.
• It translates. A deck is real text, so a viewer’s own browser can translate the talk.
• Speaker notes stay private. Notes are lifted out of the markup and never reach an audience device or a recording.
• Links open in a new tab, and a deck never runs scripts, so nothing it carries can execute on a viewer’s device.

Generating a deck is a prompt away — point an LLM at present-tense.io/llms.txt and it can write the markdown for you.

A pass over the chrome on both sides of a session.

• Present without the mouse. Pointer, draw, rectangle, clear and freeze each have a single-key shortcut, shown beside every tool in the Controls menu. They stay out of your way while you type into the page, and keyboard-driven apps like whiteboards can switch our shortcuts off.
• The presenter’s pill matches the audience’s. Same styling, one family.
• A cleaner following chip. The audience sees the presenter’s first name, shortened with an ellipsis if it’s long — and an info item in the menu now shows the session details: title, presenter, start time.
• Leaving is unmistakable. The audience’s Leave control is now a clear, labelled red button instead of a bare X that read like dismissing a toast.
• The extension popup shows its version beside the logo, so you always know what you are running.

When a presenter is moving fast, you sometimes want to stop and read. Now you can.

• Scroll freely without losing the talk. When you scroll on your own, the page stops moving under you. The content keeps updating; only the camera stops chasing the presenter.
• Come back when you’re ready. A Follow presenter button sits in one place — it never jumps around or shifts the page. One tap and you are back at the live view.
• Or just stop scrolling. Thirty seconds of stillness and you are brought back to the presenter automatically. Actively reading? The timer waits for you.
• The presenter can see the room. They see how many people are off reading at their own pace — a count only, never names — so they know to wait or move on.

The Present Tense Capture extension is now on the Chrome Web Store.

• One-click install from the store, with updates handled by Chrome automatically.
• Two ways in: the extension page carries both — the store listing, and the latest build as a zip you install manually. The zip is rebuilt with every release and can run slightly ahead of the store between releases.
• Works in Chrome and Edge, and other Chromium browsers.

The audience still installs nothing: they open your join link in any browser.

A pass on how faithfully a captured page reaches your audience.

• Images, stylesheets and fonts arrive for real. They are copied at capture time, so the audience sees the page complete, and it still renders when the original site is slow, unreachable, or behind a login. Oversized images are sent at the size they are actually shown, which makes joining faster.
• Nested scroll follows on complex apps. Dashboards and tools with scrollable panels stay with the presenter, panel by panel, not just the page.
• Highlights land on the same element for everyone. What the presenter marks is what each viewer sees marked, on their own screen.
• Smoother following. A page change no longer stutters the view while the next screen loads.

Presenting is unchanged: capture a tab, share the link, talk.

The site now shows what makes Present Tense different from a screen share, and how to see it yourself.

• Try it on your own phone first. Getting started has a step that sets the join link next to your laptop, so you can pinch into a chart label and compare it with the call app you already use.
• Point at something. A step explains how the presenter marks one exact element, and what the audience sees.
• Read it in your own language. Because the page arrives as real text, each person’s browser can translate it on their own screen. This is now a use case, a card on the homepage, and a row in the comparison table.
• New use cases: multilingual audiences, code review and PR walkthroughs, and incident reviews and ops standups, alongside all-hands, demo days and onboarding.

The homepage hero also gained a Present Tense / Screen share toggle, so the difference is visible before you ever start a session.

Present Tense is now in invite-only beta. A presenter captures a browser tab and the audience watches it as real, reflowing text — not a video.

• Capture any tab with the browser extension. The page’s structure, text, images, and styles are relayed live, and input values are masked before they leave your browser.
• Join from anywhere. The audience opens a link on a phone or laptop — no account, no app.
• A recap when it ends: duration, peak audience, and frames relayed.

We’re letting people in by email while we test in real venues. If you’d like a seat, join the beta.