01Audio in, video out
Start from the episode you already recorded.
Pick a 15–90 second excerpt, export it as MP3, WAV or M4A and pair it with the host’s photo. The video matches the clip’s length and timing; nothing in the recording is re-voiced.
Podcast video maker Turn a photo and podcast audio into video
Give an audio-only episode a visible host. Use a portrait plus a highlight clip to create social-ready podcast video.
- Set up without signing up
- Files stay local until you continue
- See the credit estimate before you render
Upload the host photo
Add the podcast clip
Set the scene
Render the episode
01Episode opener
An audio episode people can watch
A bearded host at a studio mic, made from one generated portrait and an episode intro voiced with ElevenLabs. Hedra Avatar rendered it in 16:9 — the format of a YouTube podcast upload.
- Audio-driven lip sync follows the actual episode
- Turn highlight clips into 9:16 social videos
- Use a host portrait, guest portrait, or illustrated character
- Keep the original voice and timing intact
Hedra Avatar · 16:916:9
02Show, then explain
From audio file to a show people watch.
Upload your podcast clip
Co-host · 16:916:9
02Co-host
Give the second voice a face.
Render each speaker from their own photo and audio. This co-host is a separate portrait and a separate line — “They asked their customers one question every single morning.” — rendered in 16:9 at a matching desk, so the two speakers look like one show when you alternate them.
03Clips for Shorts
Cut the highlight to vertical.
Render the host in 9:16 for Shorts, Reels and TikTok, or crop a wide episode clip to vertical in Video Editor and add a title before you post.
Video StudioAI avatar generator
Every talking-avatar tool in one place- Your starting materialA portrait and prepared audio
- The resultA visual podcast presenter
03How it works
Three steps from podcast clip to video.
Pick the on‑screen host
Upload a clear host or character portrait.
Choose the podcast excerpt
A focused 15–90 second clip works well for social distribution.
Generate the video clip
The avatar performs the original audio, ready for captions and export.
- Host photo
- Your episode audio
- 16:9 episodes
- 9:16 clips
- Captions
Finished examples
A morning briefing, a camera test, a fractions lesson
Briefing · 16:916:9
Camera test · 16:916:9
Lesson · 16:916:9
Complete AI video creation
Every episode, a video episode.
Keep the host photos in Lab and render each new excerpt as it lands. Clean the audio and make SRT captions in Audio Studio, crop and title in Video Editor, and translate the clip for other markets with DUB.
- Host and co-host from photos
- 16:9 episodes, 9:16 clips
- SRT captions from the audio
- Every clip saved in Lab
Questions before you create
Before the episode goes visual.
Does this change the podcast voice?
No. The talking-avatar workflow follows the audio file you supply.
How long should the clip be?
Short social highlights are easiest to watch and faster to render. Longer audio can be split into several clips.
Can I add captions?
Yes. Make SRT or VTT captions from the same audio in Audio Studio, or burn a short title into the clip with Video Editor.
Can I show a guest or co-host?
Yes. Render each speaker from their own photo and audio excerpt — one clip per speaker or per turn.
What is the longest clip I can render?
Up to ten minutes per video with Hedra Avatar, or 60 seconds with OmniHuman 1.5. To build an episode from several scenes, Director assembles videos of up to 3 minutes.
What does a podcast video cost?
Credits per second of the episode audio, so a longer excerpt costs more: trim it to the part you want to post before you render. There is no free plan: buy credits when you are ready, and the studio shows the estimate before you render.
Episode ready?
Let listeners see who’s talking.
Choose the host photo and the clip here — no account needed yet. Sign in when you continue: both upload to Lab and open in Talking avatar, and you see the credit estimate before anything renders.
- No camera or studio
- Original voice and timing
- 16:9 and 9:16 exports
Credit estimate before generation · Buy credits only when you need them