How to add captions to Reels: three routes that actually work

Three routes, and only one of them is a caption track

Reels captions can mean three different artefacts, and mixing them up makes the workflow harder than it needs to be. An in-app caption is generated by Instagram from the audio and placed as a sticker in the editor. A burned-in caption is text rendered into the exported video by an editor. A sidecar file is a separate SRT or VTT that a player loads beside the video. Instagram's own help page describes the first and how viewers control captions; it does not document a sidecar upload for a normal Reel.

That is not a reason to skip the SRT. A well-timed SRT is the most reliable way to get consistent text into an editor, and it is the file a translator or reviewer can work with. The decision is about the destination: if the destination is the Instagram app, the app's caption sticker or a burn-in is the practical route. If the destination is an editor, the SRT is the source of truth.

What the Instagram help page documents for Reels

The help page we checked is titled Manage reels and video captions on Instagram. It says captions may not be available to everyone, that Instagram uses speech recognition to write out the speech, and that the feature is available on the mobile apps rather than on a computer. Before sharing, after selecting the video or reel, you tap Stickers and then Captions. After sharing, the same page describes Options > Manage > Edit > Advanced settings, where closed captions can be turned on or off.

Availability is not universal, the text is machine-generated, and the control surface is the app. The page also explains how viewers choose Always show closed captions or Never show closed captions. If your account does not show the sticker, do not assume the video is broken; check the current help page and the app version before changing the file.

SRT for Reels: build the file before you choose a route

An SRT for Reels should be short-cue, plain text, and UTF-8. Generate the transcript with video to SRT, then use the SRT editor to split sentences that would occupy more than two lines on a phone. Reels often cut quickly, so the cue boundaries should follow the spoken phrase rather than the scene change. A cue that starts on the new shot but before the new sentence will feel late even when the timestamps are technically correct.

Keep a master SRT even if you plan to use the in-app sticker. The master lets you re-time the same words for a different cut, hand the text to a translator, or load it into an editor for a styled burn-in. Name the file with the project and version, for example product-demo-v3.srt, so the file you hand to an editor is not confused with an older export. Check that the last cue is not clipped and that the file ends with a newline.

1. Generate the transcript and listen to the first fifteen seconds.
2. Split any cue that runs longer than two short lines.
3. Correct names, numbers and negations against the audio.
4. Export UTF-8 SRT with sequential cue numbers.
5. Keep the master beside the project file, not only in the upload folder.

Instagram Reels caption size and safe zones

Instagram does not publish a single caption font size that applies to every Reel. The in-app sticker scales with the frame and the style you choose, while a burned-in caption is sized in the editor. That means the useful question is not what number to type but how much of the frame the text should occupy. A caption block that covers a third of the height leaves little room for the subject, especially when the interface overlays the bottom and right edges.

Use the app's preview as the authority for the sticker, and a phone-sized preview for a burn-in. Keep the text inside the central band, leave a margin from the right-side action buttons, and avoid the lowest part of the frame where the caption, username and audio line appear. Contrast is part of size too: a thin light font over a bright wall can vanish at the same height that reads well over a dark scene. If you are burning text in, a starting point of roughly thirty to thirty-five characters per line is easier to read than a full-width sentence, but check it against the actual footage rather than treating the number as a platform rule.

• Keep the caption block in the central band.
• Leave room at the bottom for the app's own text and controls.
• Leave room on the right for the action buttons.
• Test the longest name and the longest line, not the average.
• Do not place two caption styles in the same zone.

Line length and reading speed on a 9:16 frame

Vertical video compresses the reading task. A viewer holds the phone close, but the frame is narrow, the caption competes with the subject, and the interface takes space at the edges. Two lines of eight to twelve words each are usually more comfortable than three dense lines. Break at a clause boundary and keep the important word at the start of the line where the eye lands first.

Reading speed is a better guide than character count alone. If a cue contains a product name, a number and a call to action, it probably needs more time than a short reaction. Watch the clip with the sound off and with the sound on; the second pass catches cues that are technically readable but emotionally out of step with the edit. The subtitle timing guidelines explain how to balance duration against the pace of the cut.

Burning captions into the video when timing must be exact

The burn-in guide explains the export mechanics, but the Reels-specific decision is about permanence. A burn-in guarantees that the words appear in the downloaded file, in a shared link, and in any player that ignores caption tracks. It also guarantees that the words cannot be turned off or translated by the viewer. For a campaign where the look is part of the brand, that trade can be acceptable. For an interview or a tutorial, a caption track plus a clean master is more useful.

If you burn in, export a high-quality master first and create the captioned version from it. Do not stack a second compression pass on top of a file that has already been uploaded and downloaded. Keep the SRT and the project file together so a correction does not require rebuilding the entire edit from memory.

Use an SRT in an editor before you publish

An SRT becomes useful in an editor when you need typography, animation, or a consistent look across several clips. Import it into Premiere Pro, DaVinci Resolve, Final Cut Pro or CapCut, place it on the caption or subtitle track, and style the track rather than each cue. The import guide covers the menu paths and the frame-rate pitfalls for each editor.

Export a clean MP4 for Instagram and keep the SRT beside it. If the Reel is part of a larger package, the same captions can be reused for a horizontal cut, a YouTube upload, or a website player.

Accessibility and the caption controls viewers already have

Instagram already gives viewers a caption control, and the help page describes how to set it to always or never show closed captions. That control is a reason to keep a caption track or a legible burn-in rather than relying on a description alone. Captions help viewers watching without sound, viewers in noisy places, and viewers who process text more easily than speech.

Accessibility is not only about the words. Speaker changes, meaningful sounds and the difference between a pause and a cut all affect comprehension. The accessibility guide covers the wider checklist, while the closed captions vs subtitles page explains the terminology if you need to brief a client.

A publishing checklist for Reels

Before publishing, watch the Reel on a phone with the sound off, then with the sound on. Check that the first caption appears before the first word is missed, that the longest line does not collide with the interface, and that the last caption is still readable when the loop restarts. If you used the in-app sticker, review the generated text for names and numbers; speech recognition is a draft, not a proofreader.

Keep the SRT, the project file and the exported video in one folder, so a later repost does not need another transcription pass. Start from the VoiceToTextFree home page if you need the tool, and use the captioning workflow when the Reel is one part of a larger release.

• A caption track or burn-in is present and legible on a phone.
• Text avoids the bottom and right interface.
• Names, numbers and negations match the audio.
• The first and last cues survive the loop.
• The SRT master is saved with the project.