Already have the audio? Drop in a voiceover, a podcast cut, or a cloned-voice recording and ViralMint builds the short around it — word-by-word captions synced to your voice, B-roll matched to what you actually say. Local, no watermark.
No install needed — plan one in your browser →
{
"input_formats": ["mp3", "m4a", "wav", "aac", "ogg", "flac", "mp4", "mov", "webm"],
"transcription": "on-device (faster-whisper) — your audio isn't uploaded for this step",
"captions": "word-by-word, synced to your audio",
"b_roll": "royalty-free Pexels stock, auto-matched to the transcript",
"voice_source": ["your own recording", "on-device cloned voice (VoxCPM)"],
"aspect_ratios": ["9:16", "16:9", "1:1"],
"output": "mp4, no watermark",
"subscription": false,
"max_clip_minutes": 10
}Pipeline from ViralMint's open-source repo (audio_staging.py + the Smart Video pipeline), 2026-07.
Most 'audio to video' tools assume you'll type a script and let a robot read it. ViralMint flips that: upload your real recording — a talking-head voiceover, a podcast segment, a phone memo — and it becomes the video's audio track verbatim. No recording handy? Clone your own voice with the built-in VoxCPM engine, which runs 100% on your machine (the reference clip never leaves it), then narrate in it.
ViralMint transcribes your clip locally with Whisper for word-level timestamps, then burns animated word-by-word captions timed to exactly when you say each word — not an even-split approximation. Pick a caption style and aspect ratio (9:16, 16:9, or 1:1).
The transcript drives an automatic Pexels stock-footage match, scene by scene, so the visuals track your narration instead of a generic loop. Blend in your own clips, add background music, and export a clean mp4 with no watermark.
In the free ViralMint desktop app (macOS / Windows / Linux), open Smart Video and toggle the input mode from 'Write a script' to 'Use my audio'.
Upload a voiceover, podcast cut, or any clip with speech — mp3, m4a, wav, mp4, mov and more, up to about 10 minutes. ViralMint transcribes it on-device with Whisper and shows you the transcript.
Choose 9:16, 16:9, or 1:1, a caption style, and optional background music. The visuals auto-match your transcript from royalty-free Pexels stock; blend in your own clips if you want.
Click Generate. ViralMint assembles the B-roll, burns word-by-word captions synced to your audio, mixes music, and lands a watermark-free mp4 in your Library.
Last updated October 2026
| Capability | Revid.ai | ViralMint |
|---|---|---|
| Use your own recorded audio as the voice | Yes | Yes — your clip is the voice track |
| On-device voice cloning | No (cloud) | Yes — 100% local (VoxCPM) |
| Transcription runs locally (audio not uploaded) | No (cloud) | Yes — faster-whisper on your machine |
| Pricing | $39–199/mo subscription | Free app · pay-as-you-go |
| Open source | No | Yes (AGPL-3.0) |
highlighted column = clearer fit for that capability. Tied capabilities are left unmarked.
Yes — that's the whole point of this mode. Upload a recording of your own voice (or a podcast clip) and it becomes the video's audio track exactly as recorded. ViralMint transcribes it locally to sync captions and match B-roll; it does not replace your voice with a synthetic one.
Two options. Write a script and let ViralMint generate a voiceover (13 AI voices), or clone your own voice once with the built-in VoxCPM engine — it runs on your machine and the reference clip never leaves it — then narrate any script in your voice and feed that into audio-to-video.
mp3, m4a, aac, ogg, flac, wav, and video containers (mp4, mov, webm) — ViralMint extracts the audio track. Clips up to about 10 minutes; for a podcast, trim to the segment you want as a short first.
The transcription runs on your own machine via faster-whisper, so your audio isn't uploaded for that step. B-roll matching pulls royalty-free Pexels stock. The app itself is a free download for macOS, Windows, and Linux.
For an instrumental or a song with no spoken words, use the Music Visualizer instead — it turns audio into an animated waveform/bars video. Audio-to-video here is for spoken-word content, because it builds the captions from the transcript.
No. The mp4 you download is clean and yours to post anywhere — YouTube Shorts, TikTok, Reels. ViralMint has no subscription and adds no watermark.
The Audio to Video ships inside the free ViralMint desktop app — no subscription, no watermark, no per-minute cap. Download once, use on as many videos as you want.