Reels Text and Captions: What Goes On Screen and What Goes in the Caption

Most Reels are watched on mute. Nearly all of them are indexed by text. Both facts mean the words on screen are doing more work than the words in the caption, and most accounts treat them as an afterthought.

What on-screen text is for

1. Being understood without sound

If the video only makes sense with audio, a large share of the audience gets nothing. On-screen text is the accessible version of your script — not a word-for-word transcription, but enough that the idea lands silently.

2. Search indexing

Instagram indexes on-screen text. A Reel with the phrase "budget desk setup for a small room" on screen can surface in search for that phrase months after it was posted. This is the closest thing to free, permanent distribution on the platform.

Write the on-screen text the way a person would type the search. "Small room desk setup under $200" is searchable. "Obsessed with this corner" is not.

3. Holding attention

Text that changes gives the eye something to follow and signals that information is still coming. A single static line for 30 seconds communicates that nothing happens in the video.

The rules for on-screen text

• Keep it out of the UI. The bottom 20% and the right 15% are covered by the caption, profile and buttons. Text placed there is invisible.
• Five to seven words per frame. Longer than that and it is read instead of watched.
• High contrast, large enough to read on a phone. Test it by watching your own Reel on the smallest screen you own.
• Front-load the searchable phrase. Put it in the first two seconds where it is seen and indexed.
• One idea per text block. A block that says two things makes the viewer choose which to read.

What the caption is for

A different job entirely:

1. Topic signal. Instagram reads the caption to decide who to test the video on. Plain-language description of the subject beats a mood statement.
2. The part that needs space. Caveats, sources, a list of steps, the link to a longer guide.
3. The conversation. A question the audience can answer, which produces comments that carry information.
4. The credit. If the audio, the footage or the idea came from someone, say so. Reposting without credit is both an etiquette failure and a route to being reported.

What the caption is not for

• Repeating the on-screen text. The viewer already read it.
• Thirty hashtags. Three to eight accurate ones is the whole job; see hashtag strategy.
• "Link in bio." Say what is at the link instead: "The full checklist is linked in my bio."
• Apologising. "Sorry I've been quiet!" tells a new viewer nothing and makes the account look neglected.

A two-minute check before you publish

1. Watch it on mute. Does it make sense?
2. Read only the on-screen text. Does it describe what the video is about in words someone would search?
3. Read only the caption's first line. Is there a reason to tap "more"?
4. Is any text sitting under the caption bar or the buttons?
5. Does the last on-screen frame connect back to the first one?

Five questions, two minutes, and it fixes the most common reasons a good Reel reaches 1,800 people instead of 40,000.