The Three Modes at a Glance
Subflow offers three caption modes for After Effects. They all use the same underlying transcription — what differs is how that transcription gets broken into on-screen cues.
| Mode | One word per cue | Phrase-length cues | Full sentence per cue | |---|---|---|---| | Single Word | ✅ | | | | Smart Flow | | ✅ | | | Full Sentence | | | ✅ | | Best for | TikTok, Reels, Shorts | Interviews, podcasts, YouTube long-form | Documentaries, broadcast | | Reading pace | Fast (high engagement) | Medium (most readable) | Slow (calm, contemplative) |
If you only remember one rule: Smart Flow is the right default for most work, Single Word for vertical short-form, Full Sentence for slow editorial content.
Single Word — For Reels, TikTok, Viral Content
Single Word mode displays one word at a time, synced to when that word is spoken. It's the look you see on every TikTok, Instagram Reel, and YouTube Short — text aggressively pulsing on a beat, demanding attention.
This works because short-form vertical video is sound-off by default. The viewer is scrolling, half-watching, and the eye catches movement. Each new word is a new piece of motion. Your retention curve depends on holding attention through the first three seconds; rapid-fire word captions are one of the strongest tools for that.
Use Single Word when:
- Vertical 9:16 video for TikTok, Instagram Reels, YouTube Shorts, or Snapchat
- Video under 90 seconds where every second of attention matters
- The audio is fast-paced — high-energy narration or quick dialogue
- You want the captions to feel kinetic and on-trend
Avoid Single Word when:
- Video is longer than 2 minutes — viewers will fatigue
- Audio is slow, contemplative, or contains long pauses
- The platform's audience expects calmer pacing (LinkedIn, professional B2B)
Smart Flow — For Interviews, Podcasts, Educational
Smart Flow groups words into natural phrases — typically 3-7 words per cue, broken at natural speech boundaries like clause endings, pauses, or breath marks. It's the most readable mode for sustained viewing because the reader's eye can take in a meaningful phrase at a glance rather than tracking each word.
For the vast majority of long-form video — interviews, podcasts repurposed as video, tutorials, webinars, course content — Smart Flow is the default that wins on both readability and engagement metrics.
Use Smart Flow when:
- Video runs longer than 60 seconds
- The format is conversational — interviews, talking heads, podcast clips
- Audience is watching deliberately (not scrolling)
- The platform is YouTube, Vimeo, LinkedIn, an internal training portal
Avoid Smart Flow when:
- Vertical short-form social — Single Word usually outperforms
- Documentary work where the slower Full Sentence pace fits the tone better
Full Sentence — For Documentaries, Broadcast
Full Sentence puts one complete sentence per caption cue. The reader gets the whole grammatical unit, then it transitions to the next. This is the convention you see on broadcast television, in cinema subtitles, and in most professionally-produced documentary content.
It's the slowest of the three modes by design. The trade-off: lower per-second engagement, but higher comprehension and a more "serious" visual register. Documentary and broadcast viewers expect this pacing; using Single Word in those contexts would feel jarring and amateurish.
Use Full Sentence when:
- Documentary-style work or broadcast deliverables
- Content that demands a calm, contemplative tone
- Long sentences where breaking mid-phrase would lose meaning
- Multi-language subtitle work (most translators work sentence-by-sentence)
Avoid Full Sentence when:
- Short-form social — pacing will feel sluggish
- High-energy content where the rhythm of the captions should match the audio
Side-By-Side Example
The same 6 seconds of audio — "I started building Subflow because every existing tool was either too slow or too expensive" — rendered in each mode:
| Time | Single Word | Smart Flow | Full Sentence | |---|---|---|---| | 0.0s | I | I started building Subflow | I started building Subflow because every existing tool was either too slow or too expensive | | 0.3s | started | because every existing tool | | | 0.7s | building | was either too slow | | | 1.1s | Subflow | or too expensive | | | 1.5s | because | | | | 2.0s | every | | | | 2.4s | existing | | | | ... | | | |
Cue counts: Single Word produces ~14 cues, Smart Flow produces 4, Full Sentence produces 1. Same transcription, very different on-screen rhythm.
How to Switch Modes in Subflow
You don't have to commit to a mode upfront. Open the Subflow panel, select the desired mode in the dropdown, and re-generate. The underlying transcription doesn't have to re-run — Subflow caches it and just re-groups the cues, so the switch takes a couple of seconds rather than the full transcription time.
For the full workflow walkthrough, see How to Add Captions in After Effects.
Frequently Asked Questions
Can I switch modes after generating captions? Yes — switching regenerates the text layers in the new structure, but the underlying transcription is cached so it takes a couple of seconds, not the full transcription time.
Which mode is best for YouTube? Smart Flow for long-form. Single Word only for YouTube Shorts. Full Sentence for documentary-style content.
Do the modes have different accuracy? No — accuracy comes from the transcription, which is the same across all three modes. The modes only differ in how transcribed words get grouped into cues.
What to Do Next
Pick a mode based on the platform and content, generate captions, adjust per cue if needed — the layers are native AE text, fully editable. If you don't have Subflow yet, see pricing or install the plugin.