How to Turn Audio Into a Video With AI: Songs and Podcasts

Turning audio into video with AI splits into three jobs, and VdoBloom handles each with a different tool. A song becomes a shot-planned music video in Music Video Maker, a long podcast or interview becomes captioned vertical clips in Clip Studio, and a plain voiceover becomes a talking presenter video through the avatar tools. All three are driven by the audio you supply rather than by a prompt, which is the point: the timing comes from the track.

Updated October 2026.

Three kinds of audio, three routes

The reason there is no single “audio to video” button is that songs, speech and narration need different things from a video.

• A song needs a shot list that lands on the beat and a look that holds for the whole track. The video is the point.
• A podcast or interview needs the strongest moments found, cut out of a long recording, reframed and captioned. The video is a distribution format for audio you already have.
• A voiceover needs a face. If the audio is someone explaining something, a talking presenter is often the whole video.

Picking the wrong route is the most common waste of credits here. A 45-minute podcast uploaded to the music video tool will not do anything useful, and a three-minute song has no “best moments” to cut out.

Route 1: turn a song into a music video

1. Open Music Video Maker at /dashboard/video-creation/music-video-maker/ and choose the upload path rather than the generate path.
2. Upload your track. The file picker accepts mp3, wav and m4a up to 25 MB. Before it will proceed you must tick the rights box confirming you own the song or have the rights to use it in a video — this is not a formality, and the tool will not run without it.
3. Pick a length. The options are 15, 20, 30, 45 and 60 seconds. This is the length of the finished video, not the length of your audio file.
4. Choose the character. Three options: use your own photo, generate a portrait for a few credits, or describe a character in words. If you upload a photo, the consent confirmation appears and the image is capped at 15 MB.
5. Set the look, aspect and model. Aspect is 9:16 for vertical or 16:9 for landscape. The shot models come from the Seedance family, and each has its own resolution options.
6. Decide on lyric captions. The Burn lyric captions into the video checkbox adds word-timed lyrics over the footage. Leave it off if you want a clean cut.
7. Plan, then review the plan. The tool first analyses the song on its own servers to find sections and beats, then writes a shot list. You get to read that plan before anything is filmed. This is the step worth slowing down on.
8. Review the shots, then build. Draft clips come back for approval, then the final build cuts them to the beat and adds the captions.

The flow is genuinely staged: plan, shots, build, with a review gate at each one and a phase banner that tells you what is happening — making the song and planning shots, filming your shots, cutting to the beat, and finally the finished video.

Generating the song instead of uploading one

If you do not have a track, the same tool writes one. Turn the source toggle to Generate a song and you get four inputs: a prompt describing genre, mood and instruments; a music model to run it on; an instrumental toggle if you want no vocals; and a lyrics box that understands verse and chorus labels.

Three music models are offered. ACE-Step Turbo is the fast option and ACE-Step the richer one, both at 6 credits per generation. MiniMax Music is labelled the best for vocals and costs 12 credits. Worth knowing before you start: the shortest song ACE-Step produces is about 30 seconds, so if you want a 15-second video the tool trims a window out of a longer track rather than generating a shorter one.

You can also make music on its own in the audio section of the dashboard if you would rather keep generation and video planning as separate steps.

Route 2: turn a podcast into clips

Long spoken audio is a different problem. You are not shooting a video around it; you are mining it.

1. Open Clip Studio at /dashboard/video-creation/clip-studio/ and upload the recording. The limit is 90 minutes and 1 GB, and the tool says plainly that spoken content works best. There is no URL import — you upload the file.
2. Choose how many clips. Anywhere from one to five in a single run.
3. Leave captions on. The tool transcribes with word-level timing and burns word-timed captions into each clip. There are seven caption styles, from a bold spoken-word pill to a clean minimal sentence-case look, and several vertical framing layouts that decide how a wide recording is reframed for a phone screen.
4. Generate. Expect a few minutes for the cut. The tool tells you it usually takes three to six minutes.
5. Review the output. Each clip arrives with a suggested title, a hook line and a score out of ten, so you can judge which ones are worth posting rather than uploading all of them.

If you would rather caption a short clip yourself, the AI Captions tool takes a talking video of up to three minutes, transcribes it, lets you fix any word before rendering, and burns the captions in.

Route 3: a voiceover into a presenter video

Sometimes the audio is just someone explaining something, and the video is a face saying it. The avatar tools take a photo plus audio and sync the mouth to the track. If you have a script rather than a recording, Avatar Presenter converts the script to speech and drives the photo with it, and the voiceover tools in the audio section will produce a standalone track when you want one. The full set of models for photo-plus-audio work, with their caps and credit costs, is listed on the AI talking avatar page.

What it costs

Step	Credits
Generated song (ACE-Step Turbo or Base)	6
Generated song (MiniMax Music)	12
Generated character portrait	3
Each shot in a music video	Priced from the Seedance model and shot length you pick
Music video build fee	10
Clip Studio	2 credits per minute of source video, minimum 4
AI Captions	2 credits per minute of source video, minimum 4

Two things about that table are worth noting. Clip Studio and AI Captions charge by the length of the file you upload, not by how many clips come out, so five clips from an hour-long podcast costs the same as one. And the music video tool charges in stages, refunding any step that fails, so an abandoned plan does not cost you a finished video’s worth of credits. Plans run from a monthly Lite tier through to large Pro tiers, and credits can also be bought as one-time packs that never expire — the current rates are on the pricing page.

Common mistakes

• Uploading a track you do not have rights to. The music video tool asks you to confirm this for a reason. A music video you cannot publish is an expensive render.
• Ignoring the shot plan. Music Video Maker shows you the plan before it films. Most disappointing results trace back to a shot list nobody read.
• Expecting a 30-second song to fill a 60-second video. The length you choose drives the edit; a short track will not stretch.
• Captioning everything. Lyric captions suit a music video. On a clean cinematic cut they compete with the footage.
• Feeding Clip Studio a music track. It looks for speech. A song with no dialogue has nothing to cut.

If music video prompting is new to you, the patterns that hold a story together across shot changes are broken down in AI music video prompt breakdowns, and the shorter end-to-end walkthrough is in turning a song into a music video.

Frequently asked questions

Can I turn a song into a video without filming anything?

Yes. Music Video Maker plans shots from your track, generates them with the Seedance models and cuts them to the beat. You supply the song or generate one, choose a character and a look, and the tool handles the rest.

What audio formats does the music video tool accept?

mp3, wav and m4a, up to 25 MB. The rights confirmation must be ticked before the tool will run, and the finished video can be 15, 20, 30, 45 or 60 seconds long.

Can I use a podcast as the audio?

Not in the music video tool, which is built around a track and a beat grid. Long spoken recordings belong in Clip Studio, which finds the strongest moments, cuts them into 9:16 clips and burns in captions.

How many clips does Clip Studio produce?

You choose between one and five per run. Pricing is per minute of the source video with a small minimum, so the number of clips does not change what you pay.

Do I need to own the music?

You need the rights to use it. The upload path requires an explicit confirmation that you own the song or have the rights to use it in a video, and output from a paid account is watermark-free with commercial rights.

Start with the route that matches your file. Upload a track to Music Video Maker, drop a long recording into Clip Studio, or bring a voiceover to the avatar tools. New accounts get 10 free credits with no card, and every step shows its price before it runs.