Video BasicsAug 9, 2026

How to add text to video

Most on-screen text fails for one of three reasons: it sits where the app draws a button, it is too thin to read on a phone, or it leaves before anyone finishes it.

8 minute read

Satvik ShrivasFounder, FrameOS

Decide what the text is for before you style it

On-screen text does two completely different jobs and they follow different rules. Decorative text is atmosphere: a stylised word on a transition, a logotype in the corner, a graphic flourish. Meaningful text carries information the viewer needs, and if it disappears the clip stops making sense: the hook line, a price, a name, a statistic, an instruction. The accessibility standards draw the same line. WCAG 2.2 exempts pure decoration from its contrast requirement entirely, while text that conveys meaning has to clear a measured ratio. That distinction is the most useful editing rule on this page, because it tells you where to spend effort. Decoration can be as pretty as you like. Anything load-bearing has to survive a small screen, a moving background, a scrolling thumb and someone watching with the sound off.

The frame is smaller than it looks

You are not designing for a 1080 by 1920 rectangle. You are designing for whatever is left after the app has drawn its own interface on top of yours. Meta publishes a safe zone specifically for 9:16 placements in Stories and Reels, because the top strip carries the profile picture and close control and the bottom third carries the caption, the audio label and the whole column of interaction buttons. TikTok tells advertisers to keep anything important inside its UI safe zone for the same reason, and notes that the safe area itself shrinks as the post caption gets longer. That last detail matters more than any pixel value you will read in a blog post, including this one: the covered area is not fixed. Treat published numbers as a starting point, then look at your own clip in each app before it goes out.

Legibility is weight, contrast and a plate, in that order

Thin, condensed, italic and script fonts all fall apart at phone size over moving footage. Start with a heavy sans-serif at a size that looks slightly too big on your monitor, because your monitor is not the delivery device. Then deal with contrast. WCAG 2.2 asks for a ratio of at least 4.5:1 between text and background, relaxing to 3:1 for large text, which the standard defines as 18 point or 14 point bold. Caption-sized type on a phone usually counts as large, so 3:1 is the floor. The complication video adds is that your background changes: white text that clears the ratio against a dark jacket fails two seconds later against a window. A stroke helps, a drop shadow helps a little, and a solid or semi-opaque plate behind the text solves it outright by making the ratio a property of your design rather than of the footage.

Time text to speech, not to the timeline

Subtitling has spent decades measuring how fast people read, and the published limits transfer well. Netflix's English style guide caps subtitles at 42 characters per line, a maximum of two lines, and a reading speed of up to 20 characters per second for adult programmes and 17 for children's. Its general requirements set a minimum of five-sixths of a second per subtitle event and a maximum of seven seconds. Twenty characters a second works out at roughly three words a second, which is a ceiling rather than a target. Short-form is harsher still, because the viewer may be reading while deciding whether to keep watching. Cut text to one idea per card, let each card land with the words being spoken rather than a beat before or after, and give even a three-word title a full second on screen.

Where to put it: top for hooks, middle for meaning

Netflix's rule is that subtitles sit centre-justified at the top or the bottom, and move to whichever is clearer when they would otherwise collide with on-screen text. Vertical video inverts the practical version of that, because the bottom of a phone frame is the busiest real estate on the internet. A workable default: hook text high but below the top strip the app covers, captions in the middle third where the face is not, and nothing important in the bottom third at all. Two things to check on the actual clip rather than in your head. Text should not cross the speaker's mouth, because sound-off viewers read lips more than they admit. And a hook title and a caption line should never occupy the frame at the same time, since two blocks of text competing means neither gets read.

Captions are a separate system, so do not build them twice

Captions and text overlays get confused because they are both words on a video, but they are different jobs with different tooling. Captions are a timed transcript of the speech, generated automatically and corrected, and for short-form they should be burned into the pixels so they render identically on every app. Overlay text is authored: a hook, a label, a number you want to land. Build them separately and keep them out of each other's way. Two posts here cover the caption side in full, one on adding captions to clips and one on what burned-in captions are and when they beat an uploaded subtitle file, so this article leaves that ground to them. The practical takeaway is simply not to hand-type your captions as text layers. That is a transcription job, and software does it faster than you.

If the text carries meaning, it has to be reachable another way

The failure case is a clip where a number, a name or a punchline exists only as pixels. A blind viewer gets nothing. So does anyone whose feed autoplays past the frame, and so does a search engine. The fix is a scripting habit rather than a production task: say the thing out loud as well as showing it. Read the figure, name the tool, describe the before and the after instead of only cutting between them. Where that is not possible, repeat the key line in the post caption. Alt text is a different mechanism and does not apply to a video file at all, though it does apply to the thumbnail and any still image posted alongside it. The accessibility guide on this blog goes through the whole checklist with the thresholds attached.

Build the template once, then stop deciding

Almost everything above is a decision you should make exactly once. Pick the font, the weight, the stroke or plate, the two positions text is allowed to occupy, and the maximum line length, then save it and reuse it. Consistency is worth more than perfection here: text that looks the same across every clip reads as a channel, and it removes the per-clip fiddling that quietly eats an afternoon. This is also the part that scales badly by hand, which is why a batching workflow helps more than a better font. FrameOS applies one caption and text style across a whole batch of clips cut from the same recording, so the styling decision happens once rather than twenty times. 300 credits for 3 days · no card.

FAQ

Where should text go on a vertical video?

Keep it out of the top strip and the bottom third, which is where the apps draw their own interface. Meta publishes a safe zone for 9:16 Stories and Reels placements and TikTok asks advertisers to stay inside its UI safe zone, noting that the safe area shrinks as your caption gets longer. Put hook text high, captions in the middle third, and nothing important at the very bottom.

How long should text stay on screen in a video?

Long enough to read comfortably, which is longer than it feels in the edit. Netflix's timed text guidelines set a minimum of five-sixths of a second per subtitle event, a maximum of seven seconds, and a reading speed of up to 20 characters per second for adult content. For short-form, treat those as ceilings and give even a very short title a full second.

What contrast does on-screen text need?

WCAG 2.2 asks for 4.5:1 against the background, relaxing to 3:1 for large text, defined as 18 point or 14 point bold. Purely decorative text is exempt. Because video backgrounds move, the reliable approach is to sit the text on a solid or semi-opaque plate, or give it a heavy stroke, so the ratio does not depend on what is behind it.

Should I add text as captions or as an overlay?

Both, for different things. Use automatic captions for the speech, corrected and burned in so they render the same everywhere. Use overlay text only for authored elements like a hook line, a label or a figure. Do not type out the dialogue as text layers by hand — that is transcription, and software does it in seconds.

Sources

Timed Text Style Guide: General Requirements · Netflix Partner Help Center
English (USA) Timed Text Style Guide · Netflix Partner Help Center
Web Content Accessibility Guidelines (WCAG) 2.2 · W3C
About text overlays and the Safe Zone for ads in Stories and Reels · Meta Business Help Centre
Creative best practices for performance ads · TikTok Ads Manager Help

Turn one long video into a week of posts

300 credits for 3 days · no card.

Start free