Text-first videos are one of the most durable categories on TikTok. They are several formats with different jobs, and most people pick the wrong one because they treat them as one thing.
Why text works on a video platform#
- Sound-off viewing. Many people scroll muted. A format whose whole message is visible loses nothing.
- Reading is faster than listening. Speech runs at about 2.5 to 3 words per second. A reader on a short, high-contrast line goes faster and can skip. Text carries more per second than a voice.
- Low production cost. No filming, no presenter, no reshoots. That matters most on your ninetieth post.
The five formats#
1. Wall of text#
A block of writing on a plain background, held long enough to read, with music under it.
For: dense arguments and anything you have already written: a newsletter paragraph, a support reply, a blog section. Fails when: the idea is thin. The viewer finishes reading quickly and has nothing left to do. Text only beats video when the content is dense.
2. Text over a clip#
Type on top of quiet footage: a walk, a drive, a product on a table.
For: content where the words matter but you want motion. Motion buys patience on longer posts. Rule: the footage must be calm and low contrast, and related to the theme without competing with it. If the viewer wants to watch the clip, they stop reading.
3. The notes-app frame#
Text styled as a note, message or document.
For: confessions, lists, anything that should feel private. Fails when: the writing is in brand voice. The frame promises a person. Never fake a real message from a real person.
4. The stat card#
One large number with a short line of context, usually in a sequence of three or four.
For: research and comparisons where the number is the story. One number per frame. Put the source on the frame. A stat card without a source, or with a made-up number, destroys the format’s credibility. The free stat video maker builds these.
5. Captions as the design#
Spoken video where bold, word-by-word captions become the main visual. Most viewers read the captions, so they are the content you must design. See how to add captions to short videos.
Building a wall of text video#
Timing#
Work out the hold time from the words.
- Count the words on the frame.
- Divide by a comfortable phone reading pace (about 5 words per second is a safe default).
- Add 1 second for the frame to register as text.
- Add 20 percent.
As an example, a 40-word block needs about 8 seconds on screen, not the 4 a fast reader in a quiet room needs. Holding too short is the most common failure, and the writer never notices, because they already know the words. If a frame needs much more than 10 seconds, split it.
Structure#
- Claim. One short line on its own frame. This is the hook.
- Setup. Two or three lines on why the claim is not obvious.
- Body. The dense part, across one or two frames.
- Turn. One line that changes what the body meant.
- Close. One line.
Keep the layout identical across frames: same margins, size, alignment and position. The eye should not have to find the text again.
Typography#
- Short lines. Roughly 24 to 34 characters. Long lines force eye movement a phone punishes.
- Tight line height, around 1.05 to 1.2, for large display type.
- Heavy weight. Thin type disappears in a busy feed.
- Sentence case. All caps destroys word shapes, which fast reading depends on.
- One emphasis per frame. Two highlights split the eye.
- Near-white on dark. Pure white on pure black vibrates. Use a backing panel if the ground is not flat.
- Margins and safe zones. Keep clear side margins and leave the bottom of the 9:16 frame empty for the caption and interface.
Background and sound#
Use a flat colour or a soft gradient. Moving footage behind dense type slows reading. Put visual interest in the transitions instead.
Do not add a voice that reads the text aloud. It is slower than reading and defeats the format. Use music with a clear beat and cut frames on it.
On-screen text is read by machines too#
Platforms and search features extract on-screen text to work out what a video is about. So:
- Put the subject in the text, not only in the voiceover.
- Write the first frame as a standalone sentence. It is the line most likely to be extracted or screenshotted.
- Make the caption and the frame agree. Two matching signals beat one.
- Avoid words rendered only as stylised images.
Choosing a format#
| If your content is | Use |
|---|---|
| A dense argument you already wrote | Wall of text |
| Longer and needs patience | Text over a clip |
| Personal or a list | Notes frame |
| About numbers | Stat cards |
| Spoken, by you or an AI presenter | Captions as design |
When two fit, choose the cheaper to iterate. Then stick with it. A channel is recognised by its shape, so use one primary format for most posts and a second for exceptions. Put variety in what you say, not in the layout.
Do not use text to hide a thin idea#
Text videos are the fastest thing to publish when you have a slot and no idea. The result is generic advice on a nice gradient, and it teaches your audience to skip the next one on sight. Test before you post: would you send this text, alone, to one person you respect? If not, do not make it a video.
Making them with Looptake#
Describe the video in one sentence and Looptake’s AI video generator picks a viral format, including text formats, writes the script, sets captions and music, and renders 9:16. Text formats need no AI media, so they use no credits. Set typography and colours once in a brand kit and every video inherits them. For more no-camera formats, see how to make faceless videos for TikTok.