Caption your podcast in 5 minutes (Whisper + auto vertical export)

A 60-minute podcast usually has 5-10 moments worth clipping for social. The problem: finding them, transcribing them, captioning them, and exporting at 9:16 used to take an hour per clip in Premiere or DaVinci.

Vivix Caption Studiodoes the whole pipeline in ~5 minutes per clip. Here's the workflow.

What Caption Studio does

• Transcribes the entire file with Whisper — 95-98% accurate, every word timestamped to the millisecond
• Lets you scrub the timeline — click any line in the transcript to jump to that timestamp
• Drag in/out handles to set clip boundaries
• Burn captions into the video with viral-tested templates (Bold TikTok, Karaoke, Subtitle)
• Auto-crop to 9:16 with subject detection so the speaker stays centered
• Export ready-to-upload MP4 — drop into TikTok, Reels, Shorts

The 5-minute workflow

Step 1: Upload (30 seconds)

Drop an MP3, MP4, M4A, or WAV. Long-form is fine — Caption Studio transcribes the whole thing once, even a 2-hour episode.

Step 2: Whisper transcribes (1-2 minutes)

Whisper runs server-side. For a 60-minute file, it finishes in ~90 seconds. Every word gets a timestamp, so you can click any line in the transcript to scrub the playhead there.

Step 3: Pick the moment (1 minute)

Skim the transcript. When you find a clip-worthy moment (a strong quote, a controversial take, a laugh), click the line to jump there. Drag the in/out handles to set 30-90 second boundaries — that's the Goldilocks length for TikTok and Reels.

Step 4: Caption + crop (30 seconds)

Pick a caption template:

• Bold TikTok — large white text, black outline, bottom-third placement. Best for podcasts.
• Karaoke — word-by-word highlight on the active word. Works for fast-paced clips.
• Subtitle — small clean text at the bottom. Best for talking-head clips.

Caption Studio auto-crops to 9:16 vertical with subject detection. If your podcast has video of two speakers, the crop follows whichever one is talking.

Step 5: Export (30 seconds)

Click export. Caption Studio renders the captioned 9:16 MP4 with ffmpeg server-side and gives you a download link. Costs 1 clip credit per export.

Pricing

• Free signup: 1 clip credit (1 captioned export)
• Standard ($10/mo): 80 clip credits/mo
• Pro ($25/mo): 120 clip credits/mo
• Ultimate ($70/mo): 400 clip credits/mo

Clip credits are separate from the regular generation credits — they reset monthly, but unused credits roll over.

Real numbers from a real podcast

We tested with a 65-minute interview podcast. Workflow:

• Upload: 45 seconds (180 MB file)
• Whisper transcription: 2 minutes 10 seconds
• Picking 5 moments: 4 minutes (skimming the transcript)
• Captioning + exporting all 5 clips: 12 minutes
• Total: ~19 minutes for 5 publishable shorts

Same job in Premiere: ~5 hours.

Tips that earned us views

1. Pull-quotes beat play-by-play

The clips that go viral are strong claims, not slow build-ups. Look for moments where the host or guest says something opinionated — that's the hook.

2. Add a 1-second pause at the end

Drag the out-point 1 second past the end of the line. The pause lets the viewer process before scrolling. Watch-time goes up.

3. Edit the captions for clarity

Whisper gets 95-98% right but it sometimes misses uncommon names or jargon. Click any caption to fix the text — the timing stays locked to the audio.

4. Use the same template across all clips from one episode

Visual consistency = brand recognition in the algorithm. Pick Bold TikTok once, use it for every clip from that episode.

Multi-language

Whisper supports 90+ languages. The transcription works in whatever language your podcast is in — Arabic, Mandarin, Japanese, Spanish, all native quality. Captions burn in with the right script and direction (RTL for Arabic, etc.).

FAQ

Can I edit the in/out points after seeing the export?

Yes — your transcript stays in your account. Open the same project, drag new boundaries, export again (costs another clip credit).

Can I do voice-over for the clip?

Yes. Generate a voice with Vivix TTS (ElevenLabs Flash, MiniMax Speech, xAI TTS) and layer it in Caption Studio.

What about B-roll behind the audio?

Generate B-roll with any of the video models in Vivix and use it as the visual track with the podcast audio overlaid.

How accurate is Whisper on heavy accents?

~92% on heavy accents (vs. 95-98% on neutral). Errors are easy to fix — click, retype, save.

Sign up free — 1 clip credit on signup, no card, your first captioned short is free.