Meet ReadFlow — the article video, without the keyframe grind

Last month we were putting together a short explainer and needed exactly that shot: a news article on screen, four quotes lighting up as the voiceover hit them. Ten seconds of finished video. We opened the editor we already had, screenshotted the page, and started keyframing.

It took forty minutes. Not forty minutes of thinking — forty minutes of clerical work. Eleven position keyframes to get four pans that didn't lurch. A yellow rectangle drawn by hand over each phrase, each one animated from zero width with a mask, each one nudged two pixels down because the first attempt sat on the letters instead of behind them. Then the voiceover came in a beat and a half slower than we'd guessed, and every one of those keyframes had to move.

We finished it. We were not proud of it. And the whole time there was this nagging thought: none of that was creative work. The software should know where the words are.

The software should know where the words are

That's the whole idea behind ReadFlow. Instead of treating your article as a flat picture and asking you to hand-draw rectangles on top of it, ReadFlow lays the text out itself — every single word, measured and placed, with its exact position on the page recorded as it goes.

So when you drag across a phrase, the app doesn't see pixels. It sees those seven words, at those coordinates, wrapping across that line break. The highlighter knows to fatten around the descenders. The camera knows how far to zoom so the phrase fills a comfortable share of the frame. Change the font size later and everything re-fits, because nothing was ever pinned to a screenshot.

A beat is a phrase plus a camera move

ReadFlow's unit of work is the beat. Select some text, click Add Beat, and you've made one: this phrase, this highlight style, the camera move that gets there, and how long each part takes.

The camera part is the bit people underestimate. A highlight on a static page looks like a slideshow. The same highlight arriving at the end of a slow glide across the page looks like a video. ReadFlow auto-frames every beat — centred with a little headroom, zoomed proportionally to how long the phrase is, never so tight it stops reading as a page — and eases between beats with a slight settle at the end, the way a real camera operator lands a move. You can nudge any of it. Most of the time you won't want to.

There's also a handheld drift toggle. Tiny, slow, almost subliminal position noise during the holds. Turn it off and the shot is clean and corporate. Turn it on and it looks like someone is holding a camera. That one checkbox does more for the feel of the footage than any other setting in the app.

Eight ways to make a phrase matter

The classic marker swipe is the default — a translucent highlighter sweeping left to right with a slight overshoot, because real markers overshoot. But different channels have different looks, so there are eight effects in total:

• Marker swipe and underline draw — the two everyone recognises.
• Circle scribble — a hand-drawn ellipse that loops around the phrase once or twice.
• Box frame — drawn on edge by edge, good for dark-mode documents.
• Spotlight — the rest of the page dims and the phrase stays bright. Brutal and effective.
• Colour pop — the letters themselves shift to your accent colour.
• Word-by-word reveal — karaoke pacing, with the per-word interval in your hands.
• Typewriter — the phrase types itself on, caret optional.

And the page underneath can be a clean article, a newspaper, a dark doc, ruled notebook paper, aged paper, or a rounded screenshot card for the tweet-and-quote look. Pick a theme, pick a font, and the same beats work on all of them.

Timing it to your voice

Here's the feature we'd have paid for that afternoon. Record your narration first. Then hit tap-sync, play the audio out loud, and tap the spacebar each time you want the next phrase to land. That's it — the beats take their timings from your taps.

If the narration comes out at nineteen seconds and your beats add up to fourteen, type 19 into fit to duration and every beat scales proportionally, keeping the rhythm you set. No dragging eleven keyframes. No arithmetic.

The difference between a shot that took forty minutes and one that took four isn't the software being clever. It's the software knowing where the words are, so you never have to tell it twice.

Your script stays yours

ReadFlow has no accounts, no sign-up, no server. The text you paste is laid out by your own browser and rendered by your own graphics card. Projects auto-save into the browser's own storage and you can export any of them as a small JSON file to keep or hand to an editor.

That matters more than usual here. People make these videos about embargoed reporting, unreleased scripts, court filings, things they'd rather not hand to a web service in exchange for a free trial. Paste it in. It stays on your laptop.

Who we built it for

New YouTubers, mostly. The ones doing news commentary or drama recaps or study channels, who need this exact shot three times a week and don't want to learn a timeline editor to get it. It's also fine for a teacher pulling apart a paragraph, or anyone who needs a quote to look good for fifteen seconds.

Bring your own narration and your own opinions. ReadFlow just moves the camera.

ReadFlow joins InkFlow and GeoFlow as another tool you can just use here, right now, for nothing. If you make something with it we'd genuinely like to watch it — or if there's an effect you keep wishing were in the list, tell us: hello@sadapanna.com.