To turn audio into a video, clean the recording, transcribe it, and lay visuals over every line: motion scenes, B-roll or generated images, plus large captions timed to the words. Keep the exact spoken words as the script, fill the track with no blank gaps, and export in 9:16 for shorts or 16:9 for YouTube.
Real outputs from the Videotto template gallery.
Audio to video by hand vs with Otto
The manual route works, it just takes a while. In Premiere Pro, DaVinci Resolve or CapCut you import the audio, generate or type the captions, then search stock sites and drag a clip or image under every sentence. Tools like Headliner make quick audiograms (a still image, a waveform and captions), and Canva can build a slideshow over a voiceover. Descript can make a waveform video from a transcript. Each covers part of the job.
Typical time for a 60 second clip from audio
| Task | In a timeline editor | With Otto |
|---|---|---|
| Clean noise and long pauses | Roughly 5 to 10 minutes | One line in the request |
| Captions timed to the words | Roughly 10 to 15 minutes | Included |
| Find a visual for each idea | Roughly 30 to 45 minutes of searching and placing | Included |
| Music bed and levels | Roughly 5 to 10 minutes | One line in the request |
| Review and fix weak scenes | 10 minutes | 5 to 10 minutes |
These are typical estimates, and they grow fast with length: a 10 minute episode can mean 100 or more visuals placed by hand. The slow part is never the cutting. It is deciding what to show for every sentence, which is the part an AI editor takes on.
What your audio can turn into
The format depends on what kind of recording you start from. Podcasters want shorts from audio only episodes, coaches want voice notes that look like lessons, and storytellers want narration that plays like a mini documentary.
| Audio you have | Video it becomes | Typical length |
|---|---|---|
| Audio only podcast | Vertical clips with motion scenes and captions | 30 to 60 seconds per clip |
| Voice note or lesson | An explainer with keywords and simple diagrams | 45 to 90 seconds |
| Narration or script read | A documentary style short with B-roll | 60 to 120 seconds |
| Interview recording | Quote cards and a waveform with captions | 20 to 45 seconds |
| Product voiceover | A promo with product shots and callouts | 15 to 30 seconds |
Turn audio into video in 5 steps
- Clean the audio first. Ask Otto to remove background noise and long pauses. Visuals timed to a messy track end up on the wrong words.
- Upload the audio or paste a link. An audio file works, and so does a YouTube link to an episode. For long recordings, say which minutes you want.
- Pick the format and the layer. 9:16 (1080 x 1920) for shorts, 16:9 (1920 x 1080) for YouTube. Say motion scenes, B-roll, images or a mix.
- Ask for full coverage. Say "back to back, no gaps" so nothing sits on a blank screen, and "one visual per idea" so scenes last 3 to 6 seconds instead of flashing past.
- Add captions, music and review. Large captions, a quiet music bed if it suits, then watch it through once and swap any scene that misses.
Step one, done properly.
Choose the visual layer for an audio video
Most audio videos use one main layer and one accent. Decide before you start, because it changes how the whole video feels.
- Motion scenes. Animated text, icons and simple diagrams that change with each idea. Best for explainers and education.
- Stock B-roll. Real footage that matches each sentence. Best for stories, travel, business and news.
- Generated images. Pictures made for each line, with slow camera moves. Best for history, fiction and anything you cannot film.
- Kinetic captions. The words are the visual: large and bold, a few at a time. Best for motivation and quotes.
- One still with motion. A cover image with a gentle zoom, a waveform and captions. Best for full episodes, where 60 minutes of changing visuals would be overkill.
Your words stay the script. A line that comes up again and again in real requests is "do not rewrite the script, use the exact spoken words". The visuals follow the audio, never the other way round.
How footage gets matched to each sentence.
A prompt to turn audio into video
Adapted from real requests Videotto users sent to Otto. Change the colours and the layer to suit your recording.
[upload your audio or paste a YouTube link] Turn this audio into a 9:16 video. Use my exact spoken words as the script and timing. Do not rewrite or summarise. Fill the whole audio with motion scenes, back to back, no gaps. One visual per idea: keywords, simple diagrams and stock footage where it fits. Add big captions, a few words at a time. Clean look: [background colour] with one accent colour, [colour].
Shorter audio to video prompts people send
- "Fill my whole audio with motion scenes, back to back, no gaps."
- "Create a video using the audio from this YouTube link."
- "Propose visuals for this voiceover like a documentary."
- "Use the voiceover as the main script and timing reference."
- "Add B-roll for every sentence and captions on top."
Example: a 40 second voice note, beat by beat
Example, not a real customer: a personal finance coach records a 40 second voice note on her phone, about 100 words, on three habits for saving money. She uploads it with the prompt above and asks for motion scenes in navy with one yellow accent. A timeline like this is what to aim for.
Worked timeline
| Time | What she says | What is on screen |
|---|---|---|
| 0 to 3 s | "Most people save whatever is left. That is the problem." | Hook text in yellow, a wallet emptying |
| 3 to 12 s | Habit one: pay yourself first on payday | Calendar icon, an arrow moving money to a savings jar |
| 12 to 22 s | Habit two: one no spend day a week | Seven day strip, one day ticked green |
| 22 to 33 s | Habit three: cancel one subscription a month | Three app icons, one crossed out, a small count up of yearly savings |
| 33 to 40 s | "Start with one. Follow for the next three." | Her name and handle, a follow prompt |
Five scenes in 40 seconds, each 3 to 11 seconds long, with captions running across all of them. The only fix she asks for afterwards is a slower count up in the third habit so the number is readable.
Why audio videos lose viewers
- Blank gaps. A second of empty screen is where people scroll. Ask for coverage from the first frame to the last.
- Visuals that change every second. One visual per idea is calmer and easier to follow than a new picture on every word.
- On screen text that rewrites the speaker. Text that does not match the voice confuses viewers. Keep the words exact.
- Small or missing captions. With no face to watch, captions carry the video. Keep them large and in the middle third.
- Music you have no rights to. If the audio is a song or someone else's show, get permission before you post.
FAQ
How do I turn an audio only podcast into YouTube Shorts?
To turn an audio only podcast into YouTube Shorts, pick moments of 30 to 60 seconds that make sense on their own, then add visuals and large captions in a 9:16 frame. A waveform over cover art is the quickest option, while motion scenes or B-roll for each idea hold attention better. Post each moment as its own Short with a clear title.
What is an audiogram and is it enough for social media?
An audiogram is a short video of a still image, an animated waveform and captions over a piece of audio. It is quick to make and works well for sharing a full episode link or a quote. For TikTok and Reels, where viewers expect movement, audiograms often hold attention less well than videos with a new visual for each idea.
How many visuals do I need for a one minute audio video?
For a one minute audio video, plan on roughly 10 to 15 visuals, one for each idea, each on screen for about 3 to 6 seconds. Fewer than that feels static and more than that feels frantic. Captions run continuously on top, so the screen always has something to read even while a scene holds.
Can I turn a voice memo from my phone into a video?
Yes, a voice memo from your phone can become a video. Export the memo as an audio file, usually M4A on iPhone or MP3 on Android, and upload it to an editor. Record somewhere quiet with the phone close to your mouth, because a clean recording makes captions more accurate and keeps the visuals timed to the right words.
What if I only have a written script and no audio?
If you only have a written script and no audio, record yourself reading it or generate an AI voiceover first. At about 150 words a minute, a 60 second video needs a script of roughly 150 words. Once the audio exists, the same steps apply: clean it, add visuals per idea, add captions and export.
Make your next video with Otto
Tell Otto what you want in plain words. It cuts, captions and animates the video for you.
