Where to put text on an Instagram Reel without covering the building, the buttons or the captions

You added a project name and a one-line description to your Reel, posted it, and the caption bar swallowed half the words while the like button sat on top of the rest. Instagram, TikTok and YouTube Shorts all draw their interface over the bottom and right side of a vertical video, and the part of the frame that stays clear is smaller than most people expect. This guide gives the safe-zone numbers in pixels, how to keep text off the building itself, what size type reads on a phone, and when to burn text in versus relying on the platform's automatic captions.

Key takeaways

• Meta recommends keeping at least 14% of the top, 35% of the bottom and 6% of each side of a Reel free of text and logos; on a 1080 × 1920 frame that leaves a box about 950 × 979 pixels.
• TikTok doesn't publish one fixed safe zone: its ad specs say the zone depends on the video's dimensions, caption length and add-ons, and it offers downloadable templates instead.
• Architecture adds a second rule the specs don't cover: put text over sky, ground or a plain wall, never across a roofline, a corner or the main opening.
• At 1080 pixels wide, 48-pixel type shows at roughly 17 points on a typical phone; keep supporting lines at or above that and titles around 72 to 96 pixels.
• Instagram's automatic closed captions come from speech and appear at the bottom of the video, so a music-only project film gets none; any words you want seen have to be burned in.

Where to put text on an Instagram Reel

Put text inside the band that starts 14% down from the top and ends 35% up from the bottom, with at least 6% clear on each side. That is Meta's recommended safe zone for Reels, and on a 1080 × 1920 video it means keeping words between roughly 269 and 1,248 pixels from the top and 65 pixels in from each edge. Within that box, place the text over a calm area of the image, such as sky, paving or a plain wall, so it never sits on the building's key lines.

The numbers come from Meta's ad specifications for Instagram Reels, which say "consider leaving at least 14% of the top, 35% of the bottom, and 6% on each side of your asset free from text, logos, or other important creative elements" (Meta Business Help Center, as of October 2026). The page is written for ads, but organic Reels carry the same overlays: account name and caption at the bottom, action buttons on the right.

The safe box is less than half the frame, which is why text in the bottom third almost always gets covered.

What each app draws over your video

TikTok's point about caption length matters: a long post caption takes more of the frame, so a short one protects your on-screen text. Since a project film usually goes to more than one platform, design to the strictest combination: Meta's percentages plus a little extra room on the right, where the button columns run tallest. Text that is left-aligned or centered slightly left of middle survives all three. The Reels guide for architects and the TikTok guide cover the other specs that differ.

Keep text off the building

The platform specs protect your text from the interface. Nothing protects your building from your text. In an architecture video the real subject is a set of edges: the roofline against the sky, a corner, a canopy line, the opening that shows the interior. A word sitting on any of those reads as a mistake even when it is legible.

1. Find the empty part of each scene first. Most exteriors have sky or foreground paving inside the safe box; most interiors have a plain wall, a ceiling plane or a floor.
2. Let the text move between scenes. A title over the sky in the opening shot and a credit over the floor in the closing shot beats forcing both into one slot.
3. Check the whole scene, not the first frame. Camera moves drift; the sky behind your text at second one can be a facade by second four.
4. Use fewer cards. One title at the start and one line at the end usually beats a label on every scene.
5. Keep contrast against the brightest frame. White text over a pale sky fails; a soft shadow or a semi-opaque band fixes it. The accessible video guide gives the 4.5:1 contrast target.

If you storyboard the film, mark the text area next to each camera move; the six-shot storyboard template has room for it.

How big text should be on a phone

Most text in short videos is too small, because it was judged on a laptop screen. A phone shows the video at its full width, and the scale factor is easy to work out. A typical phone screen is about 390 points wide, so a 1080-pixel video is drawn at roughly 1/2.8 of its pixel size: 48-pixel type in your file appears at about 17 points on the screen, a comfortable size for a short line of reading.

Working sizes we use for a 1080 × 1920 frame:

For line length and reading time, subtitling has proper references. Netflix's English timed-text style guide limits subtitles to "42 characters per line," a "maximum two lines," and "up to 20 characters per second" for adult programs (Netflix Partner Help Center, as of October 2026). For a five-second scene that is a ceiling of about 100 characters, and marketing text should stay well under it, because the viewer is also looking at the building. If you can't read a card aloud twice before the scene ends, cut words. Avoid hairline display faces and all-caps sentences; both are slow to read on a small screen.

Burned-in text versus automatic captions

There are two kinds of words on a short video, and they solve different problems.

Burned-in text (open captions, titles, labels) is part of the pixels. It shows everywhere the file goes, and you control font, size and position, but it can't be turned off or translated.

Platform captions (closed captions) are generated by the app and drawn on top of the video. Instagram's help page explains that "when you turn automatic closed captions on, the speech in the reel is automatically written out as text at the bottom," using speech recognition (Instagram Help Center, as of October 2026). You turn them on or off for a new Reel under More options, with the "Enable closed captions" toggle, and for a posted Reel through Options, then Edit, then More options. TikTok works the same way: "captions will be automatically generated for videos that you upload," based on the video language you select on the post screen (TikTok Help Center, as of October 2026).

Two consequences follow. A project film with music and no narration gets no automatic captions at all, so anything you want read has to be burned in or written in the post caption. And if a video does have narration, leave the bottom of the safe box free for the platform's captions, then review them: names of materials and places are where speech recognition slips.

Instagram's Edits app sits between the two. Its help page describes generating captions from spoken audio, then adjusting style, opacity and placement with a "Move" control (Instagram Help Center, as of October 2026). Captions made there are exported in the file, so they behave like burned-in text.

A placement workflow that takes ten minutes

1. Make a guide overlay once: a transparent 1080 × 1920 PNG with a rectangle from 269 to 1,248 pixels down and 65 pixels in from each side. For TikTok, also keep the template from its spec page.
2. Lay the guide over the film on your editor's top track, hidden at export.
3. Place each card in that scene's calm area inside the guide, then scrub its start, middle and end for edges and contrast.
4. Export at 1080 × 1920 and preview on your phone at full and low brightness.
5. Keep the post caption short so it doesn't expand over the frame.

The cover has its own, stricter crop on the profile grid; the Reel cover guide covers it.

Adding text to an Arch2Video film

An Arch2Video film arrives as a 1080 × 1920 MP4 with no titles or captions burned in; the only optional overlay is your studio logo, and a logo film also comes in a no-logo version. The soundtrack is music rather than narration, so platform captions have nothing to transcribe; a short title and a closing line are the usual approach.

The AI director decides the order of the six five-second scenes, so place text after watching the finished film, not from your upload order. If the project is unbuilt, a small "concept rendering" label in the safe box, plus the same words in the post caption, keeps the presentation honest. For a Spanish version, the bilingual captions guide covers stacked text and two-version workflows.

Frequently asked questions

Does the Reels safe zone apply to normal posts or only to ads?

Meta publishes the 14% top, 35% bottom and 6% sides guidance in its ad specifications, but organic Reels show the same interface elements in the same places. Using the ad zone for every Reel costs nothing, and if you later boost a Reel, the text is already in the right place.

Why does Instagram cover the text at the bottom of my Reel?

The bottom of a Reel holds your account name, the post caption, the audio label and, when enabled, automatic closed captions, and Meta recommends keeping the bottom 35% of the frame free of text for that reason. On a 1080 × 1920 video, that is the lowest 672 pixels. Move the text up to sit no lower than about 1,248 pixels from the top and keep the post caption short, since a longer caption takes more room.

Should I burn captions into my video or use Instagram's automatic captions?

Use automatic captions when the video has speech; they are generated from the audio, appear at the bottom and can be switched off by viewers. Burn text in when the video is music-only, when you want a specific font and position, or when the file will travel to other platforms. On narrated videos, many studios do both: a burned-in title plus platform captions for the speech.

What font size should text be on an Instagram Reel?

On a 1080 × 1920 video, keep supporting text at about 44 pixels or larger and titles around 72 to 96 pixels. A typical phone shows a 1080-pixel video at roughly a third of its pixel size, so 48-pixel type appears at about 17 points on screen. Test on an actual phone rather than a monitor, and keep each card short enough to read twice before the scene changes.